rustfs-ecstore 1.0.1

Erasure coding storage backend for RustFS, providing efficient data storage and retrieval with redundancy.
Documentation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
// Copyright 2024 RustFS Team
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.

use super::*;
use crate::core::pools::{POOL_META_NAME, load_pool_meta_identity_observing};
use crate::disk::{self, error::DiskError};
use crate::services::rebalance::{REBAL_META_NAME, RebalStatus};
use crate::set_disk::get_lock_acquire_timeout;
use crate::storage_api_contracts::heal::HealOperations as _;
use crate::storage_api_contracts::namespace::NamespaceLocking as _;
use rustfs_common::mrf_channel::MrfDeleteMarkerPurge;
use rustfs_lock::NamespaceLockGuard;
use std::collections::BTreeSet;
use tracing::trace;

const LOG_COMPONENT_ECSTORE: &str = "ecstore";
const LOG_SUBSYSTEM_HEAL: &str = "heal";
const EVENT_HEAL_ABANDONED_PARTS: &str = "heal_abandoned_parts";
const EVENT_HEAL_FORMAT_COMPLETED: &str = "heal_format_completed";
const EVENT_HEAL_OBJECT_STARTED: &str = "heal_object_started";

/// An explicit bucket-heal admission owns one lifecycle generation through its write tail.
pub(crate) struct BucketHealScope {
    pub(crate) bucket: String,
    pub(crate) incarnation: uuid::Uuid,
    pub(crate) store: Arc<ECStore>,
    fence: super::BucketIncarnationFenceGuard,
}

impl BucketHealScope {
    pub(crate) fn check(&self) -> disk::error::Result<()> {
        if self.fence.is_lock_lost() {
            return Err(DiskError::other("bucket heal incarnation fence was lost"));
        }
        Ok(())
    }
}

tokio::task_local! {
    static BUCKET_HEAL_SCOPE: Arc<BucketHealScope>;
}

pub(crate) fn bucket_heal_scope(bucket: &str) -> Option<Arc<BucketHealScope>> {
    BUCKET_HEAL_SCOPE
        .try_with(|scope| (scope.bucket == bucket).then(|| scope.clone()))
        .ok()
        .flatten()
}

/// Resolve only the two canonical bucket records; other internal paths carry no bucket authority.
pub(super) fn bucket_metadata_owner(object: &str) -> Option<&str> {
    use crate::bucket::metadata::{BUCKET_INCARNATION_FILE, BUCKET_METADATA_FILE};
    let (prefix, rest) = object.split_once('/')?;
    let (bucket, file) = rest.split_once('/')?;
    (prefix == disk::BUCKET_META_PREFIX
        && matches!(file, BUCKET_METADATA_FILE | BUCKET_INCARNATION_FILE)
        && check_valid_bucket_name_strict(bucket).is_ok())
    .then_some(bucket)
}

pub(crate) fn bucket_heal_scope_for_object(bucket: &str, object: &str) -> Option<Arc<BucketHealScope>> {
    let owner = if bucket == RUSTFS_META_BUCKET {
        bucket_metadata_owner(object)?
    } else {
        bucket
    };
    bucket_heal_scope(owner)
}

/// Storage-owned proof for the exact version and every selected erasure location.
/// This is an in-process result, never reconstructed from admin drive telemetry.
#[derive(Debug)]
pub struct HealObjectAbsenceProof {
    pub bucket: String,
    pub object: String,
    pub version_id: String,
    pub bucket_incarnation_id: Uuid,
    pub pool_index: Option<usize>,
    pub set_index: Option<usize>,
    pub locations: Vec<(usize, usize)>,
    pub removed: bool,
}

#[derive(Debug)]
pub struct HealObjectStorageResult {
    pub item: HealResultItem,
    pub error: Option<Error>,
    pub absence: Option<HealObjectAbsenceProof>,
}

fn invalid_heal_pool_index(pool_idx: usize, pool_count: usize) -> Error {
    StorageError::InvalidArgument(
        "heal".to_string(),
        "pool".to_string(),
        format!("invalid heal pool index {pool_idx} for {pool_count} pools"),
    )
}

fn is_pool_meta_object(bucket: &str, object: &str) -> bool {
    bucket == RUSTFS_META_BUCKET && object == POOL_META_NAME
}

#[derive(Debug, Clone, Copy)]
enum HealFormatPoolSkip {
    Completed,
    Retryable,
}

fn classify_heal_format_pool(
    pool_idx: usize,
    pool_cmd_line: &str,
    pool_meta: &PoolMeta,
    rebalance_meta: Option<&RebalanceMeta>,
) -> Option<HealFormatPoolSkip> {
    let Some(pool) = pool_meta.pools.get(pool_idx) else {
        return Some(HealFormatPoolSkip::Retryable);
    };

    if pool.id != pool_idx || pool_cmd_line.is_empty() || pool.cmd_line.is_empty() || pool.cmd_line != pool_cmd_line {
        return Some(HealFormatPoolSkip::Retryable);
    }

    if let Some(decommission) = pool.decommission.as_ref() {
        if decommission.complete {
            return Some(HealFormatPoolSkip::Completed);
        }
        if decommission.failed || decommission.canceled || decommission.queued || pool_meta.is_suspended(pool_idx) {
            return Some(HealFormatPoolSkip::Retryable);
        }
    }

    if let Some(meta) = rebalance_meta {
        let Some(pool_stats) = meta.pool_stats.get(pool_idx) else {
            return Some(HealFormatPoolSkip::Retryable);
        };
        if pool_stats.info.stopping || (pool_stats.participating && pool_stats.info.status == RebalStatus::Started) {
            return Some(HealFormatPoolSkip::Retryable);
        }
    }

    None
}

fn heal_format_pool_skip_error(skip: HealFormatPoolSkip) -> Error {
    match skip {
        HealFormatPoolSkip::Completed => StorageError::NoHealRequired,
        HealFormatPoolSkip::Retryable => StorageError::SlowDown,
    }
}

fn heal_format_fence_lost_error() -> Error {
    StorageError::SlowDown
}

impl ECStore {
    pub async fn purge_delete_marker_with_proof(
        self: &Arc<Self>,
        bucket: &str,
        object: &str,
        version_id: &str,
        purge: &MrfDeleteMarkerPurge,
        opts: &HealOpts,
    ) -> Result<bool> {
        if opts.dry_run || opts.no_lock || !purge.is_valid() {
            return Err(StorageError::PreconditionFailed);
        }
        let version = Uuid::parse_str(version_id)
            .ok()
            .filter(|version| !version.is_nil())
            .ok_or(StorageError::PreconditionFailed)?;
        let (Some(pool_index), Some(set_index)) = (opts.pool, opts.set) else {
            return Err(StorageError::PreconditionFailed);
        };
        let mut marker = rustfs_filemeta::MetaDeleteMarker::default();
        let consumed = marker.unmarshal_msg(&purge.marker)?;
        if usize::try_from(consumed).ok() != Some(purge.marker.len())
            || marker.version_id != Some(version)
            || marker.stable_identity() != purge.marker_identity
        {
            return Err(StorageError::PreconditionFailed);
        }
        let marker_info = marker.clone().into_fileinfo(bucket, object, false)?;
        if !marker_info.is_canonical_delete_marker()
            || marker_info.delete_marker_incarnation() != Some(purge.marker_incarnation_id)
        {
            return Err(StorageError::PreconditionFailed);
        }

        let current = self.bucket_incarnation_id_from_disk(bucket).await?;
        if current.is_nil() {
            return Err(StorageError::BucketNotFound(bucket.to_owned()));
        }
        let original = purge.bucket_incarnation_id;
        let object = object.to_owned();
        self.run_bucket_heal_at_incarnation(bucket, current, opts, move |store, bucket, _opts| async move {
            let scope = bucket_heal_scope(&bucket).ok_or(StorageError::PreconditionFailed)?;
            scope.check()?;
            if original != current {
                let retirement_store = crate::bucket::metadata_sys::object_store_if_initialized_in(&store.ctx)
                    .await
                    .ok_or(StorageError::PreconditionFailed)?;
                if !crate::bucket::retirement::is_retired(retirement_store, &bucket, original).await? {
                    return Err(StorageError::PreconditionFailed);
                }
            }
            let pool = store
                .pools
                .get(pool_index)
                .ok_or_else(|| invalid_heal_pool_index(pool_index, store.pools.len()))?;
            let set = pool.disk_set.get(set_index).cloned().ok_or_else(|| {
                StorageError::InvalidArgument(
                    "heal".to_string(),
                    "set".to_string(),
                    format!("invalid heal set index {set_index} for pool {pool_index}"),
                )
            })?;
            set.purge_delete_marker_exact(&bucket, &object, version, marker).await
        })
        .await
    }

    pub(super) async fn run_bucket_heal_at_incarnation<T, F, Fut>(
        self: &Arc<Self>,
        bucket: &str,
        expected: uuid::Uuid,
        opts: &HealOpts,
        operation: F,
    ) -> Result<T>
    where
        T: Send + 'static,
        F: FnOnce(Arc<Self>, String, HealOpts) -> Fut + Send + 'static,
        Fut: std::future::Future<Output = Result<T>> + Send + 'static,
    {
        if expected.is_nil() || self.ctx.lock_manager().is_disabled() {
            return Err(Error::other(
                "incarnation-bound bucket heal requires a valid identity and namespace locking",
            ));
        }
        // Lock order: bucket lifecycle, then bucket/object and capacity locks used by healing.
        let fence = self.acquire_bucket_incarnation_fence(bucket, expected).await?;
        let scope = Arc::new(BucketHealScope {
            bucket: bucket.to_owned(),
            incarnation: expected,
            store: self.clone(),
            fence,
        });
        let store = self.clone();
        let bucket = bucket.to_owned();
        let opts = *opts;
        // Dropping a caller's cancellation/timeout waiter must not release the lifecycle
        // owner while a storage operation is still committing.
        tokio::spawn(BUCKET_HEAL_SCOPE.scope(scope.clone(), async move {
            scope.check().map_err(Error::from)?;
            let result = operation(store, bucket, opts).await;
            scope.check().map_err(Error::from)?;
            result
        }))
        .await
        .map_err(|error| Error::other_with_context("bucket heal owner task failed", error))?
    }

    pub async fn heal_bucket_at_incarnation(
        self: &Arc<Self>,
        bucket: &str,
        expected: uuid::Uuid,
        opts: &HealOpts,
    ) -> Result<HealResultItem> {
        self.run_bucket_heal_at_incarnation(bucket, expected, opts, |store, bucket, opts| async move {
            store.heal_bucket(&bucket, &opts).await
        })
        .await
    }

    pub async fn heal_local_bucket_at_incarnation(
        self: &Arc<Self>,
        bucket: &str,
        expected: uuid::Uuid,
        opts: &HealOpts,
        fenced_pools: Vec<usize>,
        pools: Option<Vec<usize>>,
    ) -> Result<HealResultItem> {
        self.run_bucket_heal_at_incarnation(bucket, expected, opts, |store, bucket, opts| async move {
            let peer =
                crate::cluster::rpc::peer_s3_client::LocalPeerS3Client::new_with_instance_ctx(None, pools, store.ctx.clone());
            crate::cluster::rpc::peer_s3_client::PeerS3Client::heal_bucket_with_fence(&peer, &bucket, &opts, &fenced_pools)
                .await
                .map_err(Error::from)
        })
        .await
    }

    pub async fn heal_object_at_incarnation(
        self: &Arc<Self>,
        bucket: &str,
        object: &str,
        version_id: &str,
        expected: uuid::Uuid,
        opts: &HealOpts,
    ) -> Result<HealObjectStorageResult> {
        let object = object.to_owned();
        let version_id = version_id.to_owned();
        self.run_bucket_heal_at_incarnation(bucket, expected, opts, |store, bucket, opts| async move {
            store.heal_object_with_proof(&bucket, &object, &version_id, &opts).await
        })
        .await
    }

    pub async fn heal_mrf_object_at_incarnation(
        self: &Arc<Self>,
        bucket: &str,
        object: &str,
        version_id: &str,
        expected: uuid::Uuid,
        opts: &HealOpts,
    ) -> Result<HealObjectStorageResult> {
        let object = object.to_owned();
        let version_id = version_id.to_owned();
        self.run_bucket_heal_at_incarnation(bucket, expected, opts, |store, bucket, opts| async move {
            store
                .heal_object_with_authoritative_absence_proof(&bucket, &object, &version_id, &opts)
                .await
        })
        .await
    }

    async fn acquire_heal_format_fence(
        &self,
    ) -> Result<(
        tokio::sync::MutexGuard<'_, PoolMetaWriteState>,
        NamespaceLockGuard,
        NamespaceLockGuard,
        PoolMeta,
        Option<RebalanceMeta>,
    )> {
        let metadata_pool = self
            .pools
            .first()
            .cloned()
            .ok_or_else(|| Error::other("heal format requires at least one storage pool"))?;
        let mut write_state = self.pool_meta_save_gate.lock().await;
        write_state.ensure_write_safe("heal format fence failed")?;

        // Metadata fence order is part of the decommission/rebalance protocol:
        // pool.bin must always be acquired before rebalance.bin. A read guard is
        // sufficient to freeze the pool snapshot and lets replacement format
        // recovery coexist with ordinary Heal's capacity fence. The rebalance
        // write guard still serializes format repair and excludes transitions.
        let pool_lock = metadata_pool.new_ns_lock(RUSTFS_META_BUCKET, POOL_META_NAME).await?;
        let pool_guard = pool_lock.get_read_lock(get_lock_acquire_timeout()).await?;
        let rebalance_lock = metadata_pool.new_ns_lock(RUSTFS_META_BUCKET, REBAL_META_NAME).await?;
        let rebalance_guard = rebalance_lock.get_write_lock(get_lock_acquire_timeout()).await?;

        if pool_guard.is_lock_lost() || rebalance_guard.is_lock_lost() {
            return Err(heal_format_fence_lost_error());
        }

        load_pool_meta_identity_observing(self.pools.clone(), &mut write_state).await?;
        let mut pool_meta = PoolMeta::default();
        let replica_state = pool_meta
            .load_no_lock_from_replicas_observing(self.pools.clone(), &mut write_state)
            .await?;
        write_state.observe_replicas(replica_state);
        write_state.ensure_missing_metadata_can_initialize()?;
        write_state.ensure_write_safe("heal format fence failed")?;
        if pool_meta.pools.len() != self.pools.len()
            || pool_meta.pools.iter().enumerate().any(|(pool_idx, pool)| {
                pool.id != pool_idx || pool.cmd_line.is_empty() || pool.cmd_line != self.pools[pool_idx].endpoints.cmd_line
            })
        {
            return Err(heal_format_fence_lost_error());
        }

        let mut rebalance_meta = RebalanceMeta::new();
        let rebalance_meta = match rebalance_meta
            .load_with_opts(
                metadata_pool,
                ObjectOptions {
                    no_lock: true,
                    ..Default::default()
                },
            )
            .await
        {
            Ok(()) => Some(rebalance_meta),
            Err(Error::ConfigNotFound) => None,
            Err(err) => return Err(err),
        };

        if rebalance_meta
            .as_ref()
            .is_some_and(|meta| meta.pool_stats.len() != self.pools.len())
        {
            return Err(heal_format_fence_lost_error());
        }

        write_state.ensure_write_safe("heal format fence failed")?;
        if pool_guard.is_lock_lost() || rebalance_guard.is_lock_lost() {
            return Err(heal_format_fence_lost_error());
        }

        Ok((write_state, pool_guard, rebalance_guard, pool_meta, rebalance_meta))
    }

    fn get_pools_for_heal_object(&self, opts: &HealOpts) -> Result<Vec<Arc<Sets>>> {
        match opts.pool {
            Some(pool_idx) => Ok(vec![
                self.pools
                    .get(pool_idx)
                    .cloned()
                    .ok_or_else(|| invalid_heal_pool_index(pool_idx, self.pools.len()))?,
            ]),
            None => Ok(self.pools.clone()),
        }
    }

    /// Whether a replacement set owns this pool's `pool.bin` replica.
    /// Placement is determined by the same hash as metadata writes, never by
    /// whether a shard is currently readable on the replacement disk.
    pub fn replacement_pool_metadata_required(&self, pool_index: usize, set_index: usize) -> Result<bool> {
        let pool = self
            .pools
            .get(pool_index)
            .ok_or_else(|| invalid_heal_pool_index(pool_index, self.pools.len()))?;
        let target_set = pool.get_disks_for_heal_object(
            POOL_META_NAME,
            &HealOpts {
                set: Some(set_index),
                ..Default::default()
            },
        )?;
        Ok(Arc::ptr_eq(&target_set, &pool.get_disks_by_key(POOL_META_NAME)))
    }

    /// Restore the bucket configuration and incarnation before retiring a replacement marker.
    pub async fn heal_replacement_bucket_metadata(
        self: &Arc<Self>,
        bucket: &str,
        opts: &HealOpts,
        targets: &[String],
    ) -> Result<()> {
        use crate::bucket::metadata::{BUCKET_INCARNATION_FILE, BUCKET_METADATA_FILE};
        use crate::bucket::metadata_sys::acquire_bucket_metadata_transaction_read_lock_in;

        let (Some(pool_index), Some(set_index)) = (opts.pool, opts.set) else {
            return Err(Error::PreconditionFailed);
        };
        if targets.is_empty() || opts.dry_run || opts.no_lock || check_valid_bucket_name_strict(bucket).is_err() {
            return Err(Error::PreconditionFailed);
        }
        let pool = self
            .pools
            .get(pool_index)
            .ok_or_else(|| invalid_heal_pool_index(pool_index, self.pools.len()))?;
        let target_set = pool.get_disks_for_heal_object(BUCKET_METADATA_FILE, opts)?;
        if targets.iter().any(|target| {
            target_set
                .set_endpoints
                .iter()
                .filter(|endpoint| endpoint.to_string() == *target)
                .count()
                != 1
        }) {
            return Err(Error::PreconditionFailed);
        }
        let objects: Vec<_> = [BUCKET_METADATA_FILE, BUCKET_INCARNATION_FILE]
            .into_iter()
            .map(|file| format!("{}/{bucket}/{file}", disk::BUCKET_META_PREFIX))
            .filter(|object| Arc::ptr_eq(&target_set, &pool.get_disks_by_key(object)))
            .collect();
        if objects.is_empty() {
            return Ok(());
        }

        // Read the persisted identity without lazy migration: a replacement
        // must never turn lost configuration into newly fabricated defaults.
        let incarnation = self.bucket_incarnation_id_from_disk(bucket).await?;
        let targets = targets.to_vec();
        self.run_bucket_heal_at_incarnation(bucket, incarnation, opts, move |store, bucket, opts| async move {
            // Match config writers: lifecycle, metadata transaction, object locks.
            // Keep the transaction stable through target readback, including when
            // the caller is cancelled and the storage owner finishes its write.
            let transaction = acquire_bucket_metadata_transaction_read_lock_in(&store.ctx, &bucket).await?;
            for object in objects {
                let scope = bucket_heal_scope(&bucket).ok_or(Error::PreconditionFailed)?;
                scope.check()?;
                if transaction.is_lock_lost() {
                    return Err(Error::PreconditionFailed);
                }
                // Config objects occupy one pool, unlike pool.bin. Require a
                // successful all-pool lookup before accepting another pool's ownership.
                if !store.single_pool() {
                    let (owner, _) = store
                        .get_pool_info_for_delete_marker(RUSTFS_META_BUCKET, &object, &ObjectOptions::default())
                        .await?;
                    if owner.index != pool_index {
                        continue;
                    }
                }
                let metadata_opts = HealOpts { remove: false, ..opts };
                let (result, error) = store.heal_object(RUSTFS_META_BUCKET, &object, "", &metadata_opts).await?;
                if let Some(error) = error {
                    return Err(error);
                }
                for target in &targets {
                    let mut drives = result.after.drives.iter().filter(|drive| &drive.endpoint == target);
                    if !drives.next().is_some_and(|drive| drive.state == "ok") || drives.next().is_some() {
                        return Err(Error::PreconditionFailed);
                    }
                }
                if !store
                    .replacement_targets_have_version(RUSTFS_META_BUCKET, &object, "", pool_index, set_index, &targets)
                    .await?
                {
                    return Err(Error::PreconditionFailed);
                }
            }
            if transaction.is_lock_lost() {
                return Err(Error::PreconditionFailed);
            }
            Ok(())
        })
        .await
    }

    /// Return every live erasure set selected by an object-heal scope.
    pub async fn heal_erasure_set_scopes(&self, opts: &HealOpts) -> Result<Vec<(usize, usize)>> {
        let pools = self.get_pools_for_heal_object(opts)?;
        let pool_meta = self.pool_meta.read().await;
        let mut scopes = Vec::new();

        for pool in pools {
            let suspended_complete = pool_meta.is_suspended(pool.pool_idx).then(|| {
                pool_meta
                    .pools
                    .get(pool.pool_idx)
                    .and_then(|status| status.decommission.as_ref())
                    .is_some_and(|decommission| decommission.complete)
            });
            if let Some(complete) = suspended_complete {
                if opts.pool.is_some() {
                    return Err(if complete {
                        StorageError::InvalidArgument(
                            "heal".to_string(),
                            "pool".to_string(),
                            format!("heal pool {} has completed decommission", pool.pool_idx),
                        )
                    } else {
                        Error::SlowDown
                    });
                }
                continue;
            }

            if let Some(set_idx) = opts.set {
                if set_idx >= pool.disk_set.len() {
                    return Err(StorageError::InvalidArgument(
                        "heal".to_string(),
                        "set".to_string(),
                        format!(
                            "invalid heal set index {set_idx} for pool {} with {} sets",
                            pool.pool_idx,
                            pool.disk_set.len()
                        ),
                    ));
                }
                scopes.push((pool.pool_idx, set_idx));
            } else {
                scopes.extend((0..pool.disk_set.len()).map(|set_idx| (pool.pool_idx, set_idx)));
            }
        }

        if scopes.is_empty() {
            return Err(Error::SlowDown);
        }

        Ok(scopes)
    }

    #[instrument(skip(self))]
    pub(super) async fn handle_heal_format(&self, dry_run: bool) -> Result<(HealResultItem, Option<Error>)> {
        let mut r = HealResultItem {
            heal_item_type: HealItemType::Metadata.to_string(),
            detail: "disk-format".to_string(),
            ..Default::default()
        };

        let mut count_no_heal = 0;
        let mut count_completed = 0;
        let mut first_error = None;
        for (pool_idx, pool) in self.pools.iter().enumerate() {
            let (mut write_state, pool_guard, rebalance_guard, pool_meta, rebalance_meta) =
                self.acquire_heal_format_fence().await?;
            if pool_guard.is_lock_lost() || rebalance_guard.is_lock_lost() {
                first_error.get_or_insert(heal_format_fence_lost_error());
                break;
            }
            if let Some(skip) = classify_heal_format_pool(pool_idx, &pool.endpoints.cmd_line, &pool_meta, rebalance_meta.as_ref())
            {
                if matches!(skip, HealFormatPoolSkip::Completed) {
                    count_completed += 1;
                } else {
                    first_error.get_or_insert(heal_format_pool_skip_error(skip));
                }
                continue;
            }

            let fence_lost = || {
                let lost = pool_guard.is_lock_lost()
                    || rebalance_guard.is_lock_lost()
                    || write_state.ensure_write_safe("heal format write fence failed").is_err();
                if lost {
                    write_state.block_writes_after_fence_loss();
                }
                lost
            };
            let (mut result, err) = pool.heal_format_with_fence(dry_run, fence_lost).await?;
            if let Some(err) = err {
                match err {
                    StorageError::NoHealRequired => {
                        count_no_heal += 1;
                    }
                    err => {
                        first_error.get_or_insert(err);
                    }
                }
            }
            r.disk_count += result.disk_count;
            r.set_count += result.set_count;
            r.before.drives.append(&mut result.before.drives);
            r.after.drives.append(&mut result.after.drives);

            // A lease can be lost after the final write; fail closed before
            // reporting the pool as successfully healed.
            let fence_lost = pool_guard.is_lock_lost()
                || rebalance_guard.is_lock_lost()
                || write_state.ensure_write_safe("heal format publication fence failed").is_err();
            if fence_lost {
                write_state.block_writes_after_fence_loss();
                first_error.get_or_insert(heal_format_fence_lost_error());
                break;
            }
        }
        if let Some(err) = first_error {
            return Ok((r, Some(err)));
        }
        if count_no_heal + count_completed == self.pools.len() {
            info!(
                event = EVENT_HEAL_FORMAT_COMPLETED,
                component = LOG_COMPONENT_ECSTORE,
                subsystem = LOG_SUBSYSTEM_HEAL,
                dry_run,
                result = "no_heal_required",
                pool_count = self.pools.len(),
                "Heal format completed"
            );
            return Ok((r, Some(StorageError::NoHealRequired)));
        }
        info!(
            event = EVENT_HEAL_FORMAT_COMPLETED,
            component = LOG_COMPONENT_ECSTORE,
            subsystem = LOG_SUBSYSTEM_HEAL,
            dry_run,
            result = "healed_or_inspected",
            disk_count = r.disk_count,
            set_count = r.set_count,
            "Heal format completed"
        );
        Ok((r, None))
    }

    #[instrument(skip(self, targets), fields(pool_index, set_index, target_count = targets.len()))]
    pub async fn heal_replacement_format(
        &self,
        dry_run: bool,
        pool_index: usize,
        set_index: usize,
        targets: &[String],
    ) -> Result<(HealResultItem, Option<Error>)> {
        let pool = self
            .pools
            .get(pool_index)
            .ok_or_else(|| invalid_heal_pool_index(pool_index, self.pools.len()))?;
        let set = pool.disk_set.get(set_index).cloned().ok_or_else(|| {
            StorageError::InvalidArgument(
                "heal".to_string(),
                "set".to_string(),
                format!("invalid heal set index {set_index} for pool {pool_index}"),
            )
        })?;

        let (mut write_state, pool_guard, rebalance_guard, pool_meta, rebalance_meta) = self.acquire_heal_format_fence().await?;
        if let Some(skip) = classify_heal_format_pool(pool_index, &pool.endpoints.cmd_line, &pool_meta, rebalance_meta.as_ref()) {
            return Ok((HealResultItem::default(), Some(heal_format_pool_skip_error(skip))));
        }

        let fence_lost = || {
            let lost = pool_guard.is_lock_lost()
                || rebalance_guard.is_lock_lost()
                || write_state
                    .ensure_write_safe("replacement format write fence failed")
                    .is_err();
            if lost {
                write_state.block_writes_after_fence_loss();
            }
            lost
        };
        let result = set.heal_replacement_format_with_fence(dry_run, targets, fence_lost).await?;
        let fence_lost = pool_guard.is_lock_lost()
            || rebalance_guard.is_lock_lost()
            || write_state
                .ensure_write_safe("replacement format publication fence failed")
                .is_err();
        if fence_lost {
            write_state.block_writes_after_fence_loss();
            return Ok((result.0, Some(heal_format_fence_lost_error())));
        }
        Ok(result)
    }

    #[instrument(skip(self, targets), fields(pool_index, set_index, target_count = targets.len()))]
    pub async fn replacement_targets_have_version(
        &self,
        bucket: &str,
        object: &str,
        version_id: &str,
        pool_index: usize,
        set_index: usize,
        targets: &[String],
    ) -> Result<bool> {
        let pool = self
            .pools
            .get(pool_index)
            .ok_or_else(|| invalid_heal_pool_index(pool_index, self.pools.len()))?;
        let set = pool.disk_set.get(set_index).cloned().ok_or_else(|| {
            StorageError::InvalidArgument(
                "heal".to_string(),
                "set".to_string(),
                format!("invalid heal set index {set_index} for pool {pool_index}"),
            )
        })?;

        set.replacement_targets_have_version(bucket, object, version_id, targets)
            .await
            .map_err(Into::into)
    }

    #[instrument(skip(self))]
    pub(super) async fn handle_heal_bucket(&self, bucket: &str, opts: &HealOpts) -> Result<HealResultItem> {
        let movement_gate = self.ctx.data_movement_operation_gate();
        let _movement_guard = movement_gate.read().await;
        let save_guard = self.pool_meta_save_gate.lock().await;
        save_guard.ensure_write_safe("bucket heal cannot run while pool metadata requires recovery")?;
        let mut fenced_pools = BTreeSet::new();
        {
            let pool_meta = self.pool_meta.read().await;
            fenced_pools.extend((0..pool_meta.pools.len()).filter(|pool_idx| pool_meta.is_suspended(*pool_idx)));
            if let Some(pool_idx) = opts.pool {
                if pool_idx >= pool_meta.pools.len() {
                    return Err(invalid_heal_pool_index(pool_idx, pool_meta.pools.len()));
                }
                if pool_meta.is_suspended(pool_idx) {
                    let complete = pool_meta.pools[pool_idx]
                        .decommission
                        .as_ref()
                        .is_some_and(|decommission| decommission.complete);
                    return Err(if complete {
                        StorageError::InvalidArgument(
                            "heal".to_string(),
                            "pool".to_string(),
                            format!("heal pool {pool_idx} has completed decommission"),
                        )
                    } else {
                        Error::SlowDown
                    });
                }
            }
        }

        let dispatch_fenced_pools = fenced_pools.iter().copied().collect::<Vec<_>>();
        drop(save_guard);
        let mut res = self
            .peer_sys
            .heal_bucket_with_fence_from_movement_guarded_coordinator(bucket, opts, &dispatch_fenced_pools)
            .await?;
        {
            let pool_meta = self.pool_meta.read().await;
            fenced_pools.extend((0..pool_meta.pools.len()).filter(|pool_idx| pool_meta.is_suspended(*pool_idx)));
        }
        if !fenced_pools.is_empty() {
            let pools = fenced_pools.iter().map(usize::to_string).collect::<Vec<_>>().join(", ");
            res.detail = format!("skipped: bucket-volume heal fenced on decommission-suspended pool(s): {pools}");
        }

        Ok(res)
    }

    #[instrument(level = "trace", skip(self, opts), fields(bucket = %bucket, object = %object, version_id = %version_id))]
    pub(super) async fn handle_heal_object(
        &self,
        bucket: &str,
        object: &str,
        version_id: &str,
        opts: &HealOpts,
    ) -> Result<(HealResultItem, Option<Error>)> {
        self.handle_heal_object_with_absence(
            bucket,
            object,
            version_id,
            opts,
            &mut None,
            crate::set_disk::AbsenceProofRequest {
                retirement: None,
                allow_unversioned: false,
            },
        )
        .await
    }

    pub async fn heal_object_with_proof(
        &self,
        bucket: &str,
        object: &str,
        version_id: &str,
        opts: &HealOpts,
    ) -> Result<HealObjectStorageResult> {
        self.heal_object_with_absence_proof(bucket, object, version_id, opts, false)
            .await
    }

    async fn heal_object_with_authoritative_absence_proof(
        &self,
        bucket: &str,
        object: &str,
        version_id: &str,
        opts: &HealOpts,
    ) -> Result<HealObjectStorageResult> {
        self.heal_object_with_absence_proof(bucket, object, version_id, opts, true)
            .await
    }

    async fn heal_object_with_absence_proof(
        &self,
        bucket: &str,
        object: &str,
        version_id: &str,
        opts: &HealOpts,
        allow_unversioned_absence: bool,
    ) -> Result<HealObjectStorageResult> {
        // Object heal may recreate a missing bucket volume before committing a
        // shard. Keep the current bucket generation fenced for the whole
        // operation, including the ordinary unversioned path.
        if opts.dry_run || opts.no_lock || super::utils::is_reserved_or_invalid_bucket(bucket, false) {
            let (item, error) = self.handle_heal_object(bucket, object, version_id, opts).await?;
            return Ok(HealObjectStorageResult {
                item,
                error,
                absence: None,
            });
        }

        // Match object publication: bucket lifecycle before capacity and object
        // namespace locks. Keep the incarnation pinned through proof delivery.
        // Reuse the admission's owner: reacquiring a read lock behind a queued
        // lifecycle writer would deadlock with the read lock we already hold.
        let scope = bucket_heal_scope(bucket).filter(|scope| std::ptr::eq(scope.store.as_ref(), self));
        let guard = if let Some(scope) = &scope {
            scope.check().map_err(Error::from)?;
            None
        } else {
            Some(self.acquire_bucket_lifecycle_read_lock(bucket).await?)
        };
        let fence_is_lost = || {
            scope.as_ref().is_some_and(|scope| scope.fence.is_lock_lost())
                || guard.as_ref().is_some_and(NamespaceLockGuard::is_lock_lost)
        };
        let lifecycle_guard = scope
            .as_ref()
            .and_then(|scope| scope.fence.namespace_lock_guard())
            .or(guard.as_ref())
            .ok_or_else(|| Error::other("bucket heal requires a held lifecycle guard"))?;
        let retirement = crate::bucket::retirement::MarkerRetirementContext {
            store: crate::bucket::metadata_sys::object_store_if_initialized_in(&self.ctx).await,
            current_incarnation: if let Some(scope) = &scope {
                Some(scope.incarnation)
            } else {
                self.bucket_incarnation_id_from_disk(bucket)
                    .await
                    .ok()
                    .filter(|id| !id.is_nil())
            },
            lifecycle_guard,
        };
        let mut proofs = None;
        let (item, mut error) = self
            .handle_heal_object_with_absence(
                bucket,
                object,
                version_id,
                opts,
                &mut proofs,
                crate::set_disk::AbsenceProofRequest {
                    retirement: Some(&retirement),
                    allow_unversioned: allow_unversioned_absence,
                },
            )
            .await?;
        // Read the authoritative incarnation only for an absence candidate.
        // The lifecycle guard has pinned it throughout the storage operation.
        let incarnation = if proofs.is_some() && !fence_is_lost() {
            if let Some(scope) = &scope {
                Some(scope.incarnation)
            } else {
                self.bucket_incarnation_id_from_disk(bucket)
                    .await
                    .ok()
                    .filter(|id| !id.is_nil())
            }
        } else {
            None
        };
        let absence = match (incarnation, proofs) {
            (Some(incarnation), Some(proofs)) if !fence_is_lost() => Some(HealObjectAbsenceProof {
                bucket: bucket.to_owned(),
                object: object.to_owned(),
                version_id: version_id.to_owned(),
                bucket_incarnation_id: incarnation,
                pool_index: opts.pool,
                set_index: opts.set,
                removed: proofs.iter().any(|proof| proof.removed),
                locations: proofs.into_iter().map(|proof| (proof.pool_index, proof.set_index)).collect(),
            }),
            _ => None,
        };
        if absence.is_some() {
            error = None;
        }
        Ok(HealObjectStorageResult { item, error, absence })
    }

    async fn handle_heal_object_with_absence(
        &self,
        bucket: &str,
        object: &str,
        version_id: &str,
        opts: &HealOpts,
        absence: &mut Option<Vec<crate::set_disk::HealedObjectAbsence>>,
        proof: crate::set_disk::AbsenceProofRequest<'_>,
    ) -> Result<(HealResultItem, Option<Error>)> {
        trace!(
            event = EVENT_HEAL_OBJECT_STARTED,
            component = LOG_COMPONENT_ECSTORE,
            subsystem = LOG_SUBSYSTEM_HEAL,
            bucket = %bucket,
            object = %object,
            version_id = %version_id,
            remove = opts.remove,
            scan_mode = ?opts.scan_mode,
            "Heal object started"
        );
        let object = encode_dir_object(object);

        *absence = None;
        let pools = self.get_pools_for_heal_object(opts)?;
        let requested_pool_count = pools.len();
        if let Some(set_idx) = opts.set {
            for pool in &pools {
                if set_idx >= pool.disk_set.len() {
                    let err = StorageError::InvalidArgument(
                        "heal".to_string(),
                        "set".to_string(),
                        format!(
                            "invalid heal set index {set_idx} for pool {} with {} sets",
                            pool.pool_idx,
                            pool.disk_set.len()
                        ),
                    );
                    if opts.pool.is_some() {
                        return Err(err);
                    }
                    return Ok((HealResultItem::default(), Some(err)));
                }
            }
        }
        #[cfg(test)]
        let store_id = self.id;

        let mut heal_pools = Vec::with_capacity(pools.len());
        for pool in pools.iter() {
            let suspended_complete = {
                let pool_meta = self.pool_meta.read().await;
                pool_meta.is_suspended(pool.pool_idx).then(|| {
                    pool_meta
                        .pools
                        .get(pool.pool_idx)
                        .and_then(|status| status.decommission.as_ref())
                        .is_some_and(|decommission| decommission.complete)
                })
            };
            if let Some(complete) = suspended_complete {
                if opts.pool.is_some() {
                    let _ = pool.get_disks_for_heal_object(&object, opts)?;
                    let err = if complete {
                        StorageError::InvalidArgument(
                            "heal".to_string(),
                            "pool".to_string(),
                            format!("heal pool {} has completed decommission", pool.pool_idx),
                        )
                    } else {
                        Error::SlowDown
                    };
                    return Ok((HealResultItem::default(), Some(err)));
                }
                continue;
            }
            heal_pools.push(Arc::clone(pool));
        }
        let results = if is_pool_meta_object(bucket, &object) && !opts.no_lock && !heal_pools.is_empty() {
            let target_pool_indices = heal_pools.iter().map(|pool| pool.pool_idx).collect::<Vec<_>>();
            match self.acquire_pool_meta_object_heal_fence(&target_pool_indices).await {
                Ok((pool_meta_guard, admissions)) => {
                    let fixed_set = self.pools.first().and_then(|pool| pool.disk_set.first()).cloned();
                    let futures = heal_pools.iter().zip(admissions).map(|(pool, admission)| {
                        let pool = Arc::clone(pool);
                        let pool_object = object.clone();
                        let fixed_set = fixed_set.clone();
                        let mut opts = *opts;
                        async move {
                            admission?;
                            let fixed_set = fixed_set.ok_or_else(|| Error::other("pool metadata heal requires a fixed set"))?;
                            let target_set = pool.get_disks_for_heal_object(&pool_object, &opts)?;
                            opts.no_lock = fixed_set.shares_namespace_lock_domain(&target_set).await;
                            #[cfg(test)]
                            if !opts.no_lock {
                                crate::core::pools::notify_decommission_external_heal_target_lock_attempted();
                            }
                            #[cfg(test)]
                            crate::core::pools::notify_decommission_external_heal_operation_started(store_id);
                            pool.heal_object_with_absence(bucket, &pool_object, version_id, &opts, proof)
                                .await
                        }
                    });
                    let results = join_all(futures).await;
                    drop(pool_meta_guard);
                    results
                }
                Err(err) => (0..heal_pools.len()).map(|_| Err(err.clone())).collect(),
            }
        } else {
            let mut futures = Vec::with_capacity(heal_pools.len());
            for pool in heal_pools {
                let pool_idx = pool.pool_idx;
                let pool_object = object.clone();
                let opts = *opts;
                futures.push(self.run_external_decommission_capacity_heal(
                    pool_idx,
                    bucket,
                    &object,
                    opts,
                    move |opts| async move {
                        #[cfg(test)]
                        crate::core::pools::notify_decommission_external_heal_operation_started(store_id);
                        pool.heal_object_with_absence(bucket, &pool_object, version_id, &opts, proof)
                            .await
                    },
                ));
            }
            join_all(futures).await
        };

        let mut proofs = Vec::with_capacity(requested_pool_count);
        let mut errs = Vec::with_capacity(self.pools.len());
        let mut ress = Vec::with_capacity(self.pools.len());

        for res in results.into_iter() {
            match res {
                Ok((result, err, proof)) => {
                    if let Some(proof) = proof
                        && (err.is_none()
                            || err
                                .as_ref()
                                .is_some_and(|err| is_err_object_not_found(err) || is_err_version_not_found(err)))
                    {
                        proofs.push(proof);
                    }
                    let mut result = result;
                    result.object = decode_dir_object(&result.object);
                    ress.push(result);
                    errs.push(err);
                }
                Err(err) => {
                    errs.push(Some(err));
                    ress.push(HealResultItem::default());
                }
            }
        }

        // Absence in one pool cannot discharge a responsibility covering other
        // pools, including skipped decommission sources or failed lookups.
        if requested_pool_count > 0 && proofs.len() == requested_pool_count {
            *absence = Some(proofs);
        }

        for (idx, err) in errs.iter().enumerate() {
            if err.is_none() {
                return Ok((ress.remove(idx), None));
            }
        }

        // No pool returned a nil error, return the first non 'not found' error
        for (index, err) in errs.iter().enumerate() {
            return match err {
                Some(err) => {
                    if is_err_object_not_found(err) || is_err_version_not_found(err) {
                        continue;
                    }
                    Ok((ress.remove(index), Some(err.clone())))
                }
                None => Ok((ress.remove(index), None)),
            };
        }

        // At this stage, all errors are 'not found'
        if !version_id.is_empty() {
            return Ok((HealResultItem::default(), Some(Error::FileVersionNotFound)));
        }

        Ok((HealResultItem::default(), Some(Error::FileNotFound)))
    }

    #[instrument(skip(self))]
    pub(super) async fn handle_check_abandoned_parts(&self, bucket: &str, object: &str, opts: &HealOpts) -> Result<()> {
        let object = encode_dir_object(object);
        let pools = self.get_pools_for_heal_object(opts)?;

        let mut futures = Vec::with_capacity(pools.len());
        for pool in pools.iter() {
            futures.push(pool.check_abandoned_parts(bucket, &object, opts));
        }

        let mut first_error = None;
        for result in join_all(futures).await {
            if let Err(err) = result
                && first_error.is_none()
            {
                first_error = Some(err);
            }
        }

        if let Some(err) = first_error {
            return Err(err);
        }

        trace!(
            event = EVENT_HEAL_ABANDONED_PARTS,
            component = LOG_COMPONENT_ECSTORE,
            subsystem = LOG_SUBSYSTEM_HEAL,
            state = "completed",
            result = "ok",
            bucket,
            object,
            dry_run = opts.dry_run,
            "Heal abandoned parts completed"
        );

        Ok(())
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::bucket::metadata_sys;
    use crate::cluster::rpc::PeerS3Client;
    use crate::config::com::{delete_config, read_config_no_lock_preserve_empty_with_metadata, save_config};
    use crate::core::pools::{
        DecommissionCapacityAdmission, DecommissionCapacityLockOrderBarrier, DecommissionErasureLayout,
        DecommissionPoolCapacityInfo, POOL_META_IDENTITY_NAME, PoolDecommissionInfo, PoolMetaReplicaState, PoolStatus,
        initialized_pool_meta_identity_for_test, set_decommission_capacity_info_overrides_for_test,
    };
    use crate::core::sets::HealFormatAfterSaveBarrier;
    use crate::disk::error::Result as DiskResult;
    use crate::disk::{DeleteOptions, DiskOption, DiskStore, FORMAT_CONFIG_FILE, format::FormatV3, new_disk};
    use crate::layout::endpoints::{EndpointServerPools, Endpoints, PoolEndpoints};
    use crate::runtime::instance::InstanceContext;
    use crate::services::rebalance::{
        RebalanceInfo, RebalanceStats, test_three_pool_stores_with_isolated_node_contexts, test_two_pool_stores,
    };
    use crate::storage_api_contracts::bucket::{
        BucketInfo, BucketOperations, BucketOptions, DeleteBucketOptions, MakeBucketOptions,
    };
    use crate::storage_api_contracts::object::{ObjectIO as _, ObjectOperations};
    use crate::store::init_format::{load_format_erasure, save_format_file};
    use crate::store::init_local_disks_with_instance_ctx;
    use rustfs_heal_contracts::heal_channel::DriveState;
    use tokio_util::sync::CancellationToken;

    #[derive(Debug)]
    struct BlockingHealPeer {
        started: Arc<tokio::sync::Notify>,
        release: Arc<tokio::sync::Notify>,
    }

    #[async_trait::async_trait]
    impl PeerS3Client for BlockingHealPeer {
        async fn heal_bucket(&self, _bucket: &str, _opts: &HealOpts) -> DiskResult<HealResultItem> {
            self.started.notify_one();
            self.release.notified().await;
            Ok(HealResultItem::default())
        }

        async fn make_bucket(&self, _bucket: &str, _opts: &MakeBucketOptions) -> DiskResult<()> {
            Ok(())
        }

        async fn list_bucket(&self, _opts: &BucketOptions) -> DiskResult<Vec<BucketInfo>> {
            Ok(Vec::new())
        }

        async fn delete_bucket(&self, _bucket: &str, _opts: &DeleteBucketOptions) -> DiskResult<()> {
            Ok(())
        }

        async fn get_bucket_info(&self, _bucket: &str, _opts: &BucketOptions) -> DiskResult<BucketInfo> {
            Ok(BucketInfo::default())
        }

        fn get_pools(&self) -> Option<Vec<usize>> {
            Some(vec![0, 1])
        }
    }

    #[derive(Debug)]
    struct WriterQueuedLocalHealPeer {
        movement_gate: Arc<tokio::sync::RwLock<()>>,
        writer_queued: Arc<tokio::sync::Notify>,
        writer_acquired: Arc<tokio::sync::Notify>,
    }

    #[async_trait::async_trait]
    impl PeerS3Client for WriterQueuedLocalHealPeer {
        async fn heal_bucket(&self, _bucket: &str, _opts: &HealOpts) -> DiskResult<HealResultItem> {
            let _movement_guard = self
                .movement_gate
                .try_read()
                .map_err(|_| crate::disk::error::DiskError::TooManyOpenFiles)?;
            Ok(HealResultItem::default())
        }

        async fn heal_bucket_with_fence_from_movement_guarded_coordinator(
            &self,
            _bucket: &str,
            _opts: &HealOpts,
            _fenced_pools: &[usize],
        ) -> DiskResult<HealResultItem> {
            Ok(HealResultItem::default())
        }

        async fn make_bucket(&self, _bucket: &str, _opts: &MakeBucketOptions) -> DiskResult<()> {
            Ok(())
        }

        async fn list_bucket(&self, _opts: &BucketOptions) -> DiskResult<Vec<BucketInfo>> {
            Ok(Vec::new())
        }

        async fn delete_bucket(&self, _bucket: &str, _opts: &DeleteBucketOptions) -> DiskResult<()> {
            Ok(())
        }

        async fn get_bucket_info(&self, _bucket: &str, _opts: &BucketOptions) -> DiskResult<BucketInfo> {
            let movement_gate = self.movement_gate.clone();
            let writer_acquired = self.writer_acquired.clone();
            tokio::spawn(async move {
                let _movement_guard = movement_gate.write().await;
                writer_acquired.notify_one();
            });
            while self.movement_gate.try_read().is_ok() {
                tokio::task::yield_now().await;
            }
            self.writer_queued.notify_one();
            Ok(BucketInfo::default())
        }

        fn get_pools(&self) -> Option<Vec<usize>> {
            Some(vec![0, 1])
        }
    }

    async fn minimal_heal_pool(pool_idx: usize) -> Arc<Sets> {
        let format = FormatV3::new(1, 1);
        let endpoint_url = format!("http://127.0.0.1:{}/data", 19000 + pool_idx);
        let mut endpoint = Endpoint::try_from(endpoint_url.as_str()).expect("endpoint should parse");
        endpoint.set_pool_index(pool_idx);
        endpoint.set_set_index(0);
        endpoint.set_disk_index(0);

        Sets::new(
            vec![None],
            &PoolEndpoints {
                legacy: false,
                set_count: 1,
                drives_per_set: 1,
                endpoints: Endpoints::from(vec![endpoint]),
                cmd_line: String::new(),
                platform: String::new(),
            },
            &format,
            pool_idx,
            0,
        )
        .await
        .expect("minimal pool should build")
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn bucket_incarnation_heal_preserves_successor_shards_and_allows_fresh_repair() {
        let (_root, store, shutdown) = multi_pool_heal_store().await;
        let bucket = format!("heal-incarnation-{}", Uuid::new_v4().simple());
        store
            .make_bucket(&bucket, &MakeBucketOptions::default())
            .await
            .expect("original bucket");
        let old = store
            .bucket_incarnation_id_from_disk(&bucket)
            .await
            .expect("original identity");
        store
            .delete_bucket(&bucket, &DeleteBucketOptions::default())
            .await
            .expect("normal bucket deletion");
        store
            .make_bucket(&bucket, &MakeBucketOptions::default())
            .await
            .expect("same-name successor");
        let new = store
            .bucket_incarnation_id_from_disk(&bucket)
            .await
            .expect("successor identity");
        assert_ne!(old, new);
        let object = "successor-version";
        let version = Uuid::new_v4().to_string();
        let set = store.pools[0].get_disks(0);
        let mut reader = PutObjReader::from_vec(vec![17; 512 * 1024]);
        set.put_object(
            &bucket,
            object,
            &mut reader,
            &ObjectOptions {
                versioned: true,
                version_id: Some(version.clone()),
                // Drain every rename before the fixture removes a physical shard.
                write_completion: crate::object_api::WriteCompletion::TailDrained,
                ..Default::default()
            },
        )
        .await
        .expect("write successor version");
        let missing = set.disks.read().await[0].clone().expect("missing target");
        let healthy = set.disks.read().await[1].clone().expect("healthy target");
        missing
            .delete(
                &bucket,
                object,
                DeleteOptions {
                    recursive: true,
                    immediate: true,
                    ..Default::default()
                },
            )
            .await
            .expect("remove one successor shard");
        let baseline = healthy
            .read_all(&bucket, &format!("{object}/xl.meta"))
            .await
            .expect("healthy baseline");
        let opts = HealOpts {
            pool: Some(0),
            set: Some(0),
            scan_mode: rustfs_heal_contracts::heal_channel::HealScanMode::Deep,
            ..Default::default()
        };
        // Positive control for the original bug: name-only healing repairs the successor.
        let (_, error) = store
            .heal_object(&bucket, object, &version, &opts)
            .await
            .expect("unbound control");
        assert!(error.is_none());
        assert!(missing.read_xl(&bucket, object, false).await.is_ok());
        missing
            .delete(
                &bucket,
                object,
                DeleteOptions {
                    recursive: true,
                    immediate: true,
                    ..Default::default()
                },
            )
            .await
            .expect("restore the same missing-shard condition");
        let fi = healthy
            .read_version("", &bucket, object, &version, &Default::default())
            .await
            .expect("successor metadata");
        let disk_ref = healthy.endpoint().to_string();
        assert!(
            store
                .write_local_metadata_at_incarnation(&disk_ref, (&bucket, object), fi.clone(), old)
                .await
                .is_err()
        );
        assert!(
            store
                .delete_local_path_at_incarnation(
                    &disk_ref,
                    (&bucket, object),
                    DeleteOptions {
                        recursive: true,
                        immediate: true,
                        ..Default::default()
                    },
                    old
                )
                .await
                .is_err()
        );
        assert!(
            store
                .delete_local_version_at_incarnation(
                    &disk_ref,
                    (&bucket, object),
                    fi.clone(),
                    false,
                    DeleteOptions::default(),
                    old
                )
                .await
                .is_err()
        );
        assert!(
            store
                .rename_local_data_at_incarnation(&disk_ref, (&bucket, object), &fi, (&bucket, "stale-destination"), old)
                .await
                .is_err()
        );
        for file in [".metadata.bin", ".bucket-incarnation"] {
            let metadata = format!("buckets/{bucket}/{file}");
            assert!(
                store
                    .rename_local_data_at_incarnation(&disk_ref, (&bucket, object), &fi, (RUSTFS_META_BUCKET, &metadata), old)
                    .await
                    .is_err(),
                "stale remote heal must not replace the successor's {file}"
            );
        }
        assert!(store.heal_bucket_at_incarnation(&bucket, old, &opts).await.is_err());
        assert!(
            store
                .heal_object_at_incarnation(&bucket, object, &version, old, &opts)
                .await
                .is_err()
        );
        assert!(
            missing.read_xl(&bucket, object, false).await.is_err(),
            "obsolete admission must not publish a shard"
        );
        assert_eq!(
            healthy
                .read_all(&bucket, &format!("{object}/xl.meta"))
                .await
                .expect("healthy after stale heal"),
            baseline
        );
        store
            .heal_bucket_at_incarnation(&bucket, new, &opts)
            .await
            .expect("fresh bucket metadata repair");
        let result = store
            .heal_object_at_incarnation(&bucket, object, &version, new, &opts)
            .await
            .expect("new admission repair");
        assert!(result.error.is_none(), "new admission must repair independently: {:?}", result.error);
        assert!(
            missing.read_xl(&bucket, object, false).await.is_ok(),
            "new admission publishes the missing shard"
        );
        store
            .write_local_metadata_at_incarnation(&disk_ref, (&bucket, object), fi.clone(), new)
            .await
            .expect("fresh target metadata");
        store
            .delete_local_version_at_incarnation(&disk_ref, (&bucket, object), fi, false, DeleteOptions::default(), new)
            .await
            .expect("fresh target version cleanup");
        healthy
            .write_all(&bucket, "fresh-cleanup/data", bytes::Bytes::from_static(b"temporary"))
            .await
            .expect("cleanup fixture");
        store
            .delete_local_path_at_incarnation(
                &disk_ref,
                (&bucket, "fresh-cleanup"),
                DeleteOptions {
                    recursive: true,
                    immediate: true,
                    ..Default::default()
                },
                new,
            )
            .await
            .expect("fresh target path cleanup");
        shutdown.cancel();
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn bucket_incarnation_heal_owner_survives_cancelled_waiter_and_queued_writer() {
        let (_root, store, shutdown) = multi_pool_heal_store().await;
        let bucket = format!("heal-owner-{}", Uuid::new_v4().simple());
        store
            .make_bucket(&bucket, &MakeBucketOptions::default())
            .await
            .expect("bucket");
        let identity = store.bucket_incarnation_id_from_disk(&bucket).await.expect("identity");
        let (started_tx, started_rx) = tokio::sync::oneshot::channel();
        let (release_tx, release_rx) = tokio::sync::oneshot::channel();
        let (drained_tx, drained_rx) = tokio::sync::oneshot::channel();
        let worker_store = store.clone();
        let worker_bucket = bucket.clone();
        let waiter = tokio::spawn(async move {
            worker_store
                .run_bucket_heal_at_incarnation(
                    &worker_bucket,
                    identity,
                    &HealOpts::default(),
                    move |store, bucket, opts| async move {
                        started_tx.send(()).expect("announce held generation");
                        release_rx.await.expect("release commit");
                        let result = store
                            .heal_object_with_proof(&bucket, "history.txt", &Uuid::new_v4().to_string(), &opts)
                            .await?;
                        let proof = result.absence.expect("absence proof must reuse the held lifecycle owner");
                        assert_eq!(proof.bucket_incarnation_id, identity);
                        assert_eq!(proof.locations.len(), 2);
                        drained_tx.send(()).expect("announce drain");
                        Ok(())
                    },
                )
                .await
        });
        started_rx.await.expect("operation acquired generation fence");
        waiter.abort();
        assert!(waiter.await.expect_err("waiter aborted").is_cancelled());
        let writer_store = store.clone();
        let writer_bucket = bucket.clone();
        let mut writer = Box::pin(writer_store.acquire_bucket_lifecycle_write_lock(&writer_bucket));
        assert!(
            futures::poll!(&mut writer).is_pending(),
            "cancellation must not release the live operation's generation"
        );
        release_tx.send(()).expect("finish operation");
        tokio::time::timeout(std::time::Duration::from_secs(10), drained_rx)
            .await
            .expect("proof must not reacquire a lifecycle read lock behind the queued writer")
            .expect("physical operation drains");
        let guard = tokio::time::timeout(std::time::Duration::from_secs(10), writer)
            .await
            .expect("writer resumes after drain")
            .expect("lifecycle writer");
        drop(guard);
        shutdown.cancel();
    }

    async fn minimal_heal_store() -> ECStore {
        ECStore {
            id: Uuid::new_v4(),
            disk_map: HashMap::new(),
            pools: vec![minimal_heal_pool(0).await, minimal_heal_pool(1).await],
            peer_sys: S3PeerSys {
                clients: Vec::new(),
                pools_count: 2,
            },
            pool_meta: RwLock::new(PoolMeta::default()),
            rebalance_meta: RwLock::new(None),
            decommission_cancelers: RwLock::new(Vec::new()),
            start_gate: Mutex::new(()),
            pool_meta_save_gate: Mutex::default(),
            ctx: crate::runtime::instance::bootstrap_ctx(),
            bucket_fence_registry: std::sync::Arc::default(),
        }
    }

    async fn remove_pool_meta_shard(store: &ECStore, pool_idx: usize) -> DiskStore {
        let target_set = store.pools[pool_idx].get_disks_by_key(POOL_META_NAME);
        let missing_disk = target_set.disks.read().await[0]
            .clone()
            .expect("pool metadata fixture disk should be online");
        missing_disk
            .delete(
                RUSTFS_META_BUCKET,
                POOL_META_NAME,
                DeleteOptions {
                    recursive: true,
                    immediate: true,
                    ..Default::default()
                },
            )
            .await
            .expect("one pool metadata shard should be removable");
        assert!(
            missing_disk.read_xl(RUSTFS_META_BUCKET, POOL_META_NAME, false).await.is_err(),
            "pool metadata fixture must start with one missing shard"
        );
        missing_disk
    }

    #[tokio::test]
    async fn heal_erasure_set_scopes_follow_requested_pool_and_set() {
        let store = minimal_heal_store().await;

        assert_eq!(
            store
                .heal_erasure_set_scopes(&HealOpts::default())
                .await
                .expect("unscoped heal should enumerate every live set"),
            vec![(0, 0), (1, 0)]
        );
        assert_eq!(
            store
                .heal_erasure_set_scopes(&HealOpts {
                    pool: Some(1),
                    set: Some(0),
                    ..Default::default()
                })
                .await
                .expect("scoped heal should enumerate only its requested set"),
            vec![(1, 0)]
        );

        let err = store
            .heal_erasure_set_scopes(&HealOpts {
                pool: Some(0),
                set: Some(1),
                ..Default::default()
            })
            .await
            .expect_err("an invalid set scope must fail closed");
        assert!(matches!(err, Error::InvalidArgument(..)));
    }

    fn pool_meta_with_decommission(info: PoolDecommissionInfo) -> PoolMeta {
        PoolMeta {
            pools: vec![PoolStatus {
                id: 0,
                cmd_line: "pool-0".to_string(),
                last_update: OffsetDateTime::UNIX_EPOCH,
                decommission: Some(info),
            }],
            ..Default::default()
        }
    }

    #[test]
    fn heal_format_pool_state_barriers_are_classified() {
        let active = pool_meta_with_decommission(PoolDecommissionInfo {
            start_time: Some(OffsetDateTime::UNIX_EPOCH),
            ..Default::default()
        });
        assert!(matches!(
            classify_heal_format_pool(0, "pool-0", &active, None),
            Some(HealFormatPoolSkip::Retryable)
        ));

        for info in [
            PoolDecommissionInfo {
                failed: true,
                ..Default::default()
            },
            PoolDecommissionInfo {
                canceled: true,
                ..Default::default()
            },
        ] {
            assert!(matches!(
                classify_heal_format_pool(0, "pool-0", &pool_meta_with_decommission(info), None),
                Some(HealFormatPoolSkip::Retryable)
            ));
        }

        let completed = pool_meta_with_decommission(PoolDecommissionInfo {
            complete: true,
            ..Default::default()
        });
        assert!(matches!(
            classify_heal_format_pool(0, "pool-0", &completed, None),
            Some(HealFormatPoolSkip::Completed)
        ));
    }

    #[test]
    fn heal_format_pool_rebalance_barriers_and_identity_are_fail_closed() {
        let identity_meta = pool_meta_with_decommission(PoolDecommissionInfo::default());
        let rebalance = RebalanceMeta {
            pool_stats: vec![RebalanceStats {
                participating: true,
                info: RebalanceInfo {
                    status: RebalStatus::Started,
                    ..Default::default()
                },
                ..Default::default()
            }],
            ..Default::default()
        };
        assert!(matches!(
            classify_heal_format_pool(0, "pool-0", &identity_meta, Some(&rebalance)),
            Some(HealFormatPoolSkip::Retryable)
        ));

        let stopping = RebalanceMeta {
            pool_stats: vec![RebalanceStats {
                info: RebalanceInfo {
                    stopping: true,
                    ..Default::default()
                },
                ..Default::default()
            }],
            ..Default::default()
        };
        assert!(matches!(
            classify_heal_format_pool(0, "pool-0", &identity_meta, Some(&stopping)),
            Some(HealFormatPoolSkip::Retryable)
        ));

        let identity = pool_meta_with_decommission(PoolDecommissionInfo::default());
        assert!(matches!(
            classify_heal_format_pool(0, "pool-new", &identity, None),
            Some(HealFormatPoolSkip::Retryable)
        ));

        let identity_without_decommission = PoolMeta {
            pools: vec![PoolStatus {
                id: 0,
                cmd_line: "pool-0".to_string(),
                last_update: OffsetDateTime::UNIX_EPOCH,
                decommission: None,
            }],
            ..Default::default()
        };
        assert!(matches!(
            classify_heal_format_pool(0, "pool-new", &identity_without_decommission, None),
            Some(HealFormatPoolSkip::Retryable)
        ));

        assert!(matches!(
            classify_heal_format_pool(0, "", &identity_meta, None),
            Some(HealFormatPoolSkip::Retryable)
        ));

        assert!(matches!(
            classify_heal_format_pool(0, "pool-0", &PoolMeta::default(), None),
            Some(HealFormatPoolSkip::Retryable)
        ));

        let stopped = RebalanceMeta {
            stopped_at: Some(OffsetDateTime::UNIX_EPOCH),
            pool_stats: vec![RebalanceStats {
                participating: true,
                info: RebalanceInfo {
                    status: RebalStatus::Stopped,
                    ..Default::default()
                },
                ..Default::default()
            }],
            ..Default::default()
        };
        assert!(classify_heal_format_pool(0, "pool-0", &identity_meta, Some(&stopped)).is_none());

        let stopping_after_stop = RebalanceMeta {
            stopped_at: Some(OffsetDateTime::UNIX_EPOCH),
            pool_stats: vec![RebalanceStats {
                participating: true,
                info: RebalanceInfo {
                    status: RebalStatus::Started,
                    stopping: true,
                    ..Default::default()
                },
                ..Default::default()
            }],
            ..Default::default()
        };
        assert!(matches!(
            classify_heal_format_pool(0, "pool-0", &identity_meta, Some(&stopping_after_stop)),
            Some(HealFormatPoolSkip::Retryable)
        ));
    }

    #[test]
    fn skipped_heal_format_pool_is_never_reported_as_success() {
        assert!(matches!(
            heal_format_pool_skip_error(HealFormatPoolSkip::Retryable),
            StorageError::SlowDown
        ));
        assert!(matches!(
            heal_format_pool_skip_error(HealFormatPoolSkip::Completed),
            StorageError::NoHealRequired
        ));
    }

    async fn multi_pool_heal_store() -> (tempfile::TempDir, Arc<ECStore>, CancellationToken) {
        let temp_dir = tempfile::tempdir().expect("multi-pool heal test directory should be created");
        let mut pool_endpoints = Vec::new();
        for pool_index in 0..2 {
            let mut endpoints = Vec::new();
            for disk_index in 0..4 {
                let disk_path = temp_dir.path().join(format!("pool{pool_index}-disk{disk_index}"));
                tokio::fs::create_dir_all(&disk_path)
                    .await
                    .expect("multi-pool heal test disk should be created");
                let mut endpoint = Endpoint::try_from(disk_path.to_str().expect("disk path should be utf8"))
                    .expect("test endpoint should parse");
                endpoint.set_pool_index(pool_index);
                endpoint.set_set_index(0);
                endpoint.set_disk_index(disk_index);
                endpoints.push(endpoint);
            }
            pool_endpoints.push(PoolEndpoints {
                legacy: false,
                set_count: 1,
                drives_per_set: 4,
                endpoints: Endpoints::from(endpoints),
                cmd_line: format!("heal-owner-pool-{pool_index}"),
                platform: "test".to_string(),
            });
        }

        let endpoint_pools = EndpointServerPools::from(pool_endpoints);
        let instance_ctx = Arc::new(InstanceContext::new());
        init_local_disks_with_instance_ctx(&instance_ctx, endpoint_pools.clone())
            .await
            .expect("multi-pool local disks should initialize");
        let shutdown = CancellationToken::new();
        let store = ECStore::new_with_instance_ctx(
            "127.0.0.1:0".parse().expect("test address should parse"),
            endpoint_pools,
            shutdown.clone(),
            instance_ctx,
        )
        .await
        .expect("multi-pool test store should initialize");
        metadata_sys::init_bucket_metadata_sys(store.clone(), Vec::new()).await;
        (temp_dir, store, shutdown)
    }

    #[test]
    fn bucket_metadata_heal_owner_rejects_path_aliases_and_unrelated_records() {
        for file in [".metadata.bin", ".bucket-incarnation"] {
            assert_eq!(bucket_metadata_owner(&format!("buckets/example/{file}")), Some("example"));
        }
        for path in [
            "/buckets/example/.metadata.bin",
            "buckets/../.metadata.bin",
            "buckets/example/../.metadata.bin",
            "buckets/example/.metadata.bin/extra",
            "buckets/example//.metadata.bin",
            "buckets/example/.metadata.bin.old",
            "config/iam/example/.metadata.bin",
            "buckets/example/usage.json",
        ] {
            assert_eq!(bucket_metadata_owner(path), None, "unexpected bucket authority for {path}");
        }
    }

    #[tokio::test]
    async fn replacement_bucket_metadata_respects_pool_ownership_and_required_records() {
        let (root, store, shutdown) = multi_pool_heal_store().await;
        let bucket = "replacement-bucket-metadata";
        for invalid_bucket in ["MixedCase", "../example", RUSTFS_META_BUCKET] {
            assert!(matches!(
                store
                    .heal_replacement_bucket_metadata(
                        invalid_bucket,
                        &HealOpts {
                            pool: Some(0),
                            set: Some(0),
                            ..Default::default()
                        },
                        &[root.path().join("pool0-disk0").to_string_lossy().into_owned()],
                    )
                    .await,
                Err(Error::PreconditionFailed)
            ));
        }
        store
            .make_bucket(bucket, &MakeBucketOptions::default())
            .await
            .expect("create bucket and its persisted records");
        let mut owned = 0;
        for pool in 0..2 {
            let target = root.path().join(format!("pool{pool}-disk0"));
            let records: Vec<_> = [".metadata.bin", ".bucket-incarnation"]
                .into_iter()
                .map(|file| {
                    let path = target.join(RUSTFS_META_BUCKET).join("buckets").join(bucket).join(file);
                    let exists = path.join("xl.meta").exists();
                    if exists {
                        std::fs::remove_dir_all(&path).expect("remove replacement metadata shard");
                        owned += 1;
                    }
                    (path, exists)
                })
                .collect();
            store
                .heal_replacement_bucket_metadata(
                    bucket,
                    &HealOpts {
                        pool: Some(pool),
                        set: Some(0),
                        recreate: true,
                        ..Default::default()
                    },
                    &[target.to_string_lossy().into_owned()],
                )
                .await
                .expect("repair owned records and accept authoritative ownership in the other pool");
            for (path, existed) in records {
                assert_eq!(path.join("xl.meta").exists(), existed, "wrong placement for {}", path.display());
            }
        }
        assert_eq!(owned, 2, "each required record must have exactly one owning pool");
        let object = format!("buckets/{bucket}/.metadata.bin");
        let owner = store
            .get_pool_idx_existing_with_opts(RUSTFS_META_BUCKET, &object, &ObjectOptions::default())
            .await
            .expect("persisted metadata owner");
        let disk = store.pools[owner].disk_set[0].disks.read().await[0]
            .clone()
            .expect("metadata target disk");
        let fi = disk
            .read_version(
                "",
                RUSTFS_META_BUCKET,
                &object,
                "",
                &disk::ReadOptions {
                    read_data: true,
                    ..Default::default()
                },
            )
            .await
            .expect("inline metadata to repair through target-side admission");
        assert!(fi.data.is_some(), "fixture must contain an inline metadata shard");
        let target_path = root.path().join(format!("pool{owner}-disk0"));
        std::fs::remove_dir_all(target_path.join(RUSTFS_META_BUCKET).join(&object)).expect("remove remote target shard");
        let incarnation = store
            .bucket_incarnation_id_from_disk(bucket)
            .await
            .expect("current incarnation");
        store
            .rename_local_data_at_incarnation(
                &disk.endpoint().to_string(),
                (disk::RUSTFS_META_TMP_BUCKET, "metadata-heal-control"),
                &fi,
                (RUSTFS_META_BUCKET, &object),
                incarnation,
            )
            .await
            .expect("the target-side RPC admission must accept current bucket metadata repair");
        assert!(target_path.join(RUSTFS_META_BUCKET).join(&object).join("xl.meta").exists());
        let (configuration, _) = read_config_no_lock_preserve_empty_with_metadata(store.clone(), &object)
            .await
            .expect("read the current bucket configuration");
        save_config(store.clone(), &object, configuration)
            .await
            .expect("publish a newer config revision in the same bucket incarnation");
        let committed =
            std::fs::read(target_path.join(RUSTFS_META_BUCKET).join(&object).join("xl.meta")).expect("new target configuration");
        assert!(
            store
                .rename_local_data_at_incarnation(
                    &disk.endpoint().to_string(),
                    (disk::RUSTFS_META_TMP_BUCKET, "delayed-metadata-heal"),
                    &fi,
                    (RUSTFS_META_BUCKET, &object),
                    incarnation,
                )
                .await
                .is_err(),
            "a delayed repair must not overwrite a newer config in the same incarnation"
        );
        assert_eq!(
            std::fs::read(target_path.join(RUSTFS_META_BUCKET).join(&object).join("xl.meta")).expect("retained configuration"),
            committed,
            "rejected repair must leave the newer configuration unchanged"
        );
        for pool in 0..2 {
            for disk in 0..4 {
                let path = root
                    .path()
                    .join(format!("pool{pool}-disk{disk}"))
                    .join(RUSTFS_META_BUCKET)
                    .join("buckets")
                    .join(bucket)
                    .join(".metadata.bin");
                if path.exists() {
                    std::fs::remove_dir_all(path).expect("remove every copy of required bucket configuration");
                }
            }
        }
        metadata_sys::remove_bucket_metadata_in(&store.ctx, bucket)
            .await
            .expect("remove the cached configuration so repair cannot rely on it");
        assert!(
            store
                .heal_replacement_bucket_metadata(
                    bucket,
                    &HealOpts {
                        pool: Some(0),
                        set: Some(0),
                        recreate: true,
                        ..Default::default()
                    },
                    &[root.path().join("pool0-disk0").to_string_lossy().into_owned()],
                )
                .await
                .is_err(),
            "total metadata loss must not be accepted as ownership in another pool"
        );
        for pool in 0..2 {
            assert!(
                !root
                    .path()
                    .join(format!("pool{pool}-disk0"))
                    .join(RUSTFS_META_BUCKET)
                    .join(&object)
                    .exists(),
                "missing required configuration must not be recreated with defaults"
            );
        }
        shutdown.cancel();
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn absence_proof_requires_every_selected_pool() {
        let (_temp_dir, store, shutdown) = multi_pool_heal_store().await;
        let bucket = format!("absence-scope-{}", Uuid::new_v4().simple());
        store
            .make_bucket(&bucket, &MakeBucketOptions::default())
            .await
            .expect("create bucket in both pools");
        let version = Uuid::new_v4().to_string();
        let result = store
            .heal_object_with_proof(&bucket, "history.txt", &version, &HealOpts::default())
            .await
            .expect("exact absence lookup should complete across both pools");
        assert!(result.error.is_none());
        let proof = result.absence.expect("all selected pools proved the exact version absent");
        assert_eq!(proof.locations, vec![(0, 0), (1, 0)]);
        assert_eq!((proof.pool_index, proof.set_index), (None, None));
        assert_eq!(proof.bucket, bucket);
        assert_eq!(proof.version_id, version);
        assert!(!proof.removed, "already absent versions do not count as another cleanup");

        store.pool_meta.write().await.pools[1].decommission = Some(PoolDecommissionInfo {
            start_time: Some(OffsetDateTime::now_utc()),
            ..Default::default()
        });
        let partial = store
            .heal_object_with_proof(&bucket, "history.txt", &version, &HealOpts::default())
            .await
            .expect("unscoped heal should retain its legacy suspended-pool behavior");
        assert!(partial.absence.is_none(), "an uninspected suspended pool prevents global absence proof");
        shutdown.cancel();
    }

    fn heal_test_format_path(temp_dir: &tempfile::TempDir, pool_index: usize, disk_index: usize) -> std::path::PathBuf {
        temp_dir
            .path()
            .join(format!("pool{pool_index}-disk{disk_index}"))
            .join(crate::disk::RUSTFS_META_BUCKET)
            .join(FORMAT_CONFIG_FILE)
    }

    async fn remove_heal_test_format(
        temp_dir: &tempfile::TempDir,
        store: &ECStore,
        pool_index: usize,
        disk_index: usize,
    ) -> String {
        let endpoint = store.pools[pool_index].endpoints.endpoints.as_ref()[disk_index].clone();
        let target = endpoint.to_string();
        let format_path = heal_test_format_path(temp_dir, pool_index, disk_index);
        tokio::fs::remove_file(&format_path)
            .await
            .expect("replacement target format should be removable");
        assert!(
            !tokio::fs::try_exists(&format_path)
                .await
                .expect("replacement target format path should be inspectable")
        );
        let replacement = new_disk(
            &endpoint,
            &DiskOption {
                cleanup: false,
                health_check: false,
            },
        )
        .await
        .expect("unformatted replacement disk should open");
        let set = &store.pools[pool_index].disk_set[0];
        let target_slot = set
            .set_endpoints
            .iter()
            .position(|candidate| candidate == &endpoint)
            .expect("replacement endpoint should belong to its set");
        set.disks.write().await[target_slot] = Some(replacement);
        target
    }

    async fn assert_heal_test_format_missing(temp_dir: &tempfile::TempDir, pool_index: usize, disk_index: usize) {
        assert!(
            !tokio::fs::try_exists(heal_test_format_path(temp_dir, pool_index, disk_index))
                .await
                .expect("replacement target format path should be inspectable"),
            "format heal must not write the replacement target"
        );
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn replacement_format_heal_coexists_with_capacity_read_fence() {
        let (temp_dir, store, shutdown) = multi_pool_heal_store().await;
        let target = remove_heal_test_format(&temp_dir, &store, 0, 3).await;
        let capacity_guard = store
            .acquire_external_decommission_capacity_fence(&[0], DecommissionCapacityAdmission::Heal)
            .await
            .expect("ordinary heal capacity fence should be acquired");

        let (_, err) =
            tokio::time::timeout(std::time::Duration::from_secs(2), store.heal_replacement_format(false, 0, 0, &[target]))
                .await
                .expect("replacement format heal must not wait on an existing capacity read fence")
                .expect("replacement format heal should complete");

        assert!(err.is_none(), "replacement format heal should succeed: {err:?}");
        assert!(
            tokio::fs::try_exists(heal_test_format_path(&temp_dir, 0, 3))
                .await
                .expect("replacement target format path should be inspectable"),
            "replacement format heal should restore the target while the capacity read fence is held"
        );
        drop(capacity_guard);
        shutdown.cancel();
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn replacement_format_heal_remains_blocked_by_pool_metadata_writer() {
        let (temp_dir, store, shutdown) = multi_pool_heal_store().await;
        let target = remove_heal_test_format(&temp_dir, &store, 0, 3).await;
        let pool_lock = store.pools[0]
            .new_ns_lock(RUSTFS_META_BUCKET, POOL_META_NAME)
            .await
            .expect("pool metadata lock should be created");
        let pool_writer = pool_lock
            .get_write_lock(get_lock_acquire_timeout())
            .await
            .expect("pool metadata writer should be acquired");

        assert!(
            tokio::time::timeout(
                std::time::Duration::from_secs(1),
                store.heal_replacement_format(false, 0, 0, std::slice::from_ref(&target)),
            )
            .await
            .is_err(),
            "replacement format heal must wait for a pool metadata writer"
        );
        assert_heal_test_format_missing(&temp_dir, 0, 3).await;

        drop(pool_writer);
        let (_, err) =
            tokio::time::timeout(std::time::Duration::from_secs(30), store.heal_replacement_format(false, 0, 0, &[target]))
                .await
                .expect("replacement format heal should resume after the writer releases")
                .expect("replacement format heal should complete after the writer releases");
        assert!(err.is_none(), "replacement format heal should succeed: {err:?}");
        assert!(
            tokio::fs::try_exists(heal_test_format_path(&temp_dir, 0, 3))
                .await
                .expect("replacement target format path should be inspectable"),
            "replacement format heal should restore the target after the writer releases"
        );
        shutdown.cancel();
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn full_format_heal_preblocked_pool_metadata_never_writes_format() {
        let (temp_dir, store, shutdown) = multi_pool_heal_store().await;
        remove_heal_test_format(&temp_dir, &store, 0, 3).await;
        store.pool_meta_save_gate.lock().await.observe_replicas(PoolMetaReplicaState {
            needs_repair: true,
            repair_write_safe: false,
        });

        let err = store
            .handle_heal_format(false)
            .await
            .expect_err("a preblocked pool metadata state must reject full format heal");
        assert!(
            err.to_string()
                .contains("restart after all replicas are readable and consistent")
        );
        assert_heal_test_format_missing(&temp_dir, 0, 3).await;
        store
            .ensure_pool_meta_side_effects_safe("preblocked format heal side effect")
            .await
            .expect_err("the preblocked state must remain sticky");
        shutdown.cancel();
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn full_format_heal_future_identity_never_writes_format_and_latches() {
        let (temp_dir, store, shutdown) = multi_pool_heal_store().await;
        remove_heal_test_format(&temp_dir, &store, 0, 3).await;
        let (mut future_identity, _) =
            read_config_no_lock_preserve_empty_with_metadata(store.pools[1].clone(), POOL_META_IDENTITY_NAME)
                .await
                .expect("current identity should be readable");
        let future_version = u16::from_le_bytes([future_identity[2], future_identity[3]])
            .checked_add(1)
            .expect("identity version should have a future value");
        future_identity[2..4].copy_from_slice(&future_version.to_le_bytes());
        save_config(store.pools[1].clone(), POOL_META_IDENTITY_NAME, future_identity)
            .await
            .expect("future identity should be persisted");

        let err = store
            .handle_heal_format(false)
            .await
            .expect_err("a future identity must reject full format heal");
        assert!(err.to_string().contains("pool metadata incompatible"));
        assert_heal_test_format_missing(&temp_dir, 0, 3).await;
        store
            .ensure_pool_meta_side_effects_safe("future identity format heal side effect")
            .await
            .expect_err("the future identity rejection must latch the write gate");
        shutdown.cancel();
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn replacement_format_heal_epoch_conflict_never_writes_format_and_latches() {
        let (temp_dir, store, shutdown) = multi_pool_heal_store().await;
        let target = remove_heal_test_format(&temp_dir, &store, 0, 3).await;
        let conflicting_identity =
            initialized_pool_meta_identity_for_test(store.id, 2).expect("conflicting identity should encode");
        save_config(store.pools[1].clone(), POOL_META_IDENTITY_NAME, conflicting_identity)
            .await
            .expect("conflicting identity should be persisted");

        let err = store
            .heal_replacement_format(false, 0, 0, &[target])
            .await
            .expect_err("an identity epoch conflict must reject replacement format heal");
        assert!(
            err.to_string()
                .contains("identity replicas disagree on cluster identity or epoch")
        );
        assert_heal_test_format_missing(&temp_dir, 0, 3).await;
        store
            .ensure_pool_meta_side_effects_safe("epoch conflict replacement format side effect")
            .await
            .expect_err("the identity epoch conflict must latch the write gate");
        shutdown.cancel();
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn replacement_format_heal_initialized_identity_without_pool_meta_never_writes_and_latches() {
        let (temp_dir, store, shutdown) = multi_pool_heal_store().await;
        let target = remove_heal_test_format(&temp_dir, &store, 0, 3).await;
        for pool in &store.pools {
            delete_config(pool.clone(), POOL_META_NAME)
                .await
                .expect("pool metadata replica should be removable");
        }

        let err = store
            .heal_replacement_format(false, 0, 0, &[target])
            .await
            .expect_err("initialized identity without pool metadata must reject replacement format heal");
        assert!(err.to_string().contains("initialized cluster identity exists"));
        assert_heal_test_format_missing(&temp_dir, 0, 3).await;
        store
            .ensure_pool_meta_side_effects_safe("missing pool metadata replacement format side effect")
            .await
            .expect_err("missing initialized pool metadata must latch the write gate");
        shutdown.cancel();
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn full_format_heal_lost_after_last_save_does_not_publish_or_renew_and_latches() {
        let (temp_dir, store, shutdown) = multi_pool_heal_store().await;
        let target = remove_heal_test_format(&temp_dir, &store, 0, 3).await;
        store.pools[0].disk_set[0].disks.write().await[3] = None;
        let barrier = HealFormatAfterSaveBarrier::install(&store.pools[0], 3);
        let recovery_latch = store.pool_meta_save_gate.lock().await.aborted_transaction_latch_for_test();
        let mut heal = tokio::spawn({
            let store = Arc::clone(&store);
            async move { store.handle_heal_format(false).await }
        });

        barrier.wait_until_paused().await;
        let saved = tokio::fs::read(heal_test_format_path(&temp_dir, 0, 3))
            .await
            .expect("the last replacement format must be durable before fence loss");
        let saved = FormatV3::try_from(saved.as_slice()).expect("the durable replacement format should decode");
        assert_eq!(saved.erasure.this, store.pools[0].format.erasure.sets[0][3]);
        recovery_latch.store(true, std::sync::atomic::Ordering::SeqCst);
        barrier.release();

        let (result, err) = tokio::time::timeout(std::time::Duration::from_secs(30), &mut heal)
            .await
            .expect("format heal should stop after the lost fence")
            .expect("format heal task should not panic")
            .expect("format heal should return its fenced result");
        assert!(matches!(err, Some(StorageError::SlowDown)));
        assert!(
            result
                .after
                .drives
                .iter()
                .any(|drive| drive.endpoint == target && drive.state == DriveState::Missing.to_string()),
            "the lost fence must prevent the durable format from being published in the heal result"
        );
        assert!(
            store.pools[0].disk_set[0].disks.read().await[3].is_none(),
            "the lost fence must prevent renew_disk from attaching the replacement"
        );
        recovery_latch.store(false, std::sync::atomic::Ordering::SeqCst);
        store
            .ensure_pool_meta_side_effects_safe("post-save format fence loss side effect")
            .await
            .expect_err("post-save format fence loss must remain sticky");
        shutdown.cancel();
    }

    #[tokio::test]
    async fn heal_object_pool_scope_selects_only_requested_pool() {
        let store = minimal_heal_store().await;
        let pools = store
            .get_pools_for_heal_object(&HealOpts {
                pool: Some(1),
                ..Default::default()
            })
            .expect("requested pool should be selected");

        assert_eq!(pools.len(), 1);
        assert!(Arc::ptr_eq(&pools[0], &store.pools[1]));
    }

    #[tokio::test]
    async fn heal_object_pool_scope_rejects_invalid_pool() {
        let store = minimal_heal_store().await;
        let err = store
            .get_pools_for_heal_object(&HealOpts {
                pool: Some(2),
                ..Default::default()
            })
            .expect_err("out-of-range pool scope must fail closed");

        assert!(
            matches!(err, StorageError::InvalidArgument(_, ref field, ref reason)
                if field == "pool" && reason.contains("invalid heal pool index 2 for 2 pools")),
            "unexpected invalid pool error: {err:?}"
        );
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn pool_meta_read_repair_reuses_its_write_fence() {
        let (_temp_dirs, store, _other_store) = test_two_pool_stores(None).await;
        let missing_disk = remove_pool_meta_shard(&store, 0).await;

        let (result, err) = tokio::time::timeout(
            std::time::Duration::from_secs(30),
            store.handle_heal_object(
                RUSTFS_META_BUCKET,
                POOL_META_NAME,
                "",
                &HealOpts {
                    read_repair: true,
                    pool: Some(0),
                    ..Default::default()
                },
            ),
        )
        .await
        .expect("pool metadata read repair must not wait on its own capacity fence")
        .expect("pool metadata read repair should complete");

        assert!(err.is_none(), "pool metadata read repair should succeed: {err:?}");
        assert_eq!(result.object, POOL_META_NAME);
        assert!(
            missing_disk.read_xl(RUSTFS_META_BUCKET, POOL_META_NAME, false).await.is_ok(),
            "pool metadata read repair should restore the missing shard"
        );
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn pool_meta_heal_preserves_typed_lock_timeout() {
        let (_temp_dirs, store, _other_store) = test_two_pool_stores(None).await;
        let lock = store.pools[0]
            .new_ns_lock(RUSTFS_META_BUCKET, POOL_META_NAME)
            .await
            .expect("pool metadata lock should be created");
        let guard = lock
            .get_write_lock(get_lock_acquire_timeout())
            .await
            .expect("pool metadata lock should be acquired");

        let (_, err) = temp_env::async_with_vars(
            [(rustfs_config::ENV_OBJECT_LOCK_ACQUIRE_TIMEOUT, Some("1"))],
            store.handle_heal_object(
                RUSTFS_META_BUCKET,
                POOL_META_NAME,
                "",
                &HealOpts {
                    pool: Some(0),
                    ..Default::default()
                },
            ),
        )
        .await
        .expect("pool metadata lock timeout should be mapped into the heal result");

        assert!(
            matches!(err, Some(Error::Lock(rustfs_lock::LockError::Timeout { .. }))),
            "pool metadata heal must preserve the recoverable lock timeout: {err:?}"
        );
        drop(guard);
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn pool_meta_neighbor_keeps_ordinary_object_locking() {
        let (_temp_dirs, store, _other_store) = test_two_pool_stores(None).await;
        let object = "pool.bin.backup";
        save_config(store.pools[0].clone(), object, b"neighbor metadata".to_vec())
            .await
            .expect("neighbor metadata fixture should be written");

        let target_set = store.pools[0].get_disks_by_key(object);
        let lock = target_set
            .new_ns_lock(RUSTFS_META_BUCKET, object)
            .await
            .expect("neighbor metadata lock should be created");
        let guard = lock
            .get_write_lock(get_lock_acquire_timeout())
            .await
            .expect("neighbor metadata lock should be acquired");
        let barrier = DecommissionCapacityLockOrderBarrier::install(store.id, store.id);
        let heal_store = Arc::clone(&store);
        let mut heal = tokio::spawn(async move {
            heal_store
                .handle_heal_object(
                    RUSTFS_META_BUCKET,
                    object,
                    "",
                    &HealOpts {
                        pool: Some(0),
                        ..Default::default()
                    },
                )
                .await
        });

        barrier.wait_until_external_heal_target_lock_attempted().await;
        tokio::task::yield_now().await;
        assert!(!heal.is_finished(), "neighbor metadata heal must retain ordinary object locking");
        drop(guard);

        let (result, err) = tokio::time::timeout(std::time::Duration::from_secs(30), &mut heal)
            .await
            .expect("neighbor metadata heal should finish after the object lock is released")
            .expect("neighbor metadata heal task should not panic")
            .expect("neighbor metadata heal should complete");
        assert!(err.is_none(), "neighbor metadata heal should succeed: {err:?}");
        assert_eq!(result.object, object);
        drop(barrier);
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn targeted_heal_is_blocked_by_exact_fit_decommission_reservation() {
        let (_temp_dirs, store, _other_store) = test_two_pool_stores(None).await;
        let bucket = format!("heal-capacity-{}", Uuid::new_v4().simple());
        let object = "targeted-heal.bin";
        store
            .make_bucket(&bucket, &MakeBucketOptions::default())
            .await
            .expect("heal capacity bucket should be created");

        let target_set = store.pools[1].get_disks(0);
        let mut reader = PutObjReader::from_vec(b"targeted heal body".to_vec());
        target_set
            .put_object(&bucket, object, &mut reader, &ObjectOptions::default())
            .await
            .expect("targeted heal fixture should be written");
        let missing_disk = target_set.disks.read().await[0]
            .clone()
            .expect("targeted heal fixture disk should be online");
        missing_disk
            .delete(
                &bucket,
                object,
                DeleteOptions {
                    recursive: true,
                    immediate: true,
                    ..Default::default()
                },
            )
            .await
            .expect("targeted heal fixture shard should be removed");
        assert!(
            missing_disk.read_xl(&bucket, object, false).await.is_err(),
            "targeted heal fixture must start with one missing metadata copy"
        );

        let layout = DecommissionErasureLayout { data: 1, parity: 0 };
        set_decommission_capacity_info_overrides_for_test(
            store.id,
            vec![vec![
                DecommissionPoolCapacityInfo::for_test(0, layout, 0, 30, 30),
                DecommissionPoolCapacityInfo::for_test(1, layout, 60, 60, 0),
            ]],
        );
        store
            .save_current_pool_meta_for_decommission_start(&[0], Vec::new())
            .await
            .expect("the exact-fit decommission reservation should activate");
        {
            let pool_meta = store.pool_meta.read().await;
            let reservation = pool_meta.pools[0]
                .decommission
                .as_ref()
                .and_then(|info| info.capacity_reservation.as_ref())
                .expect("the active decommission reservation should be durable");
            assert_eq!(reservation.targets.len(), 1);
            assert_eq!(reservation.targets[0].pool_index, 1);
            assert_eq!(reservation.targets[0].reserved_physical_bytes, 60);
        }

        let (_, err) = store
            .handle_heal_object(
                &bucket,
                object,
                "",
                &HealOpts {
                    pool: Some(1),
                    set: Some(0),
                    ..Default::default()
                },
            )
            .await
            .expect("targeted heal should return a mapped capacity result");
        assert!(matches!(err, Some(Error::SlowDown)), "targeted heal must be capacity-blocked: {err:?}");
        assert!(
            missing_disk.read_xl(&bucket, object, false).await.is_err(),
            "capacity-blocked targeted heal must not rewrite the missing shard"
        );

        let (_, err) = tokio::time::timeout(
            std::time::Duration::from_secs(30),
            store.handle_heal_object(
                RUSTFS_META_BUCKET,
                POOL_META_NAME,
                "",
                &HealOpts {
                    pool: Some(1),
                    ..Default::default()
                },
            ),
        )
        .await
        .expect("pool metadata admission should not recurse on its namespace lock")
        .expect("pool metadata capacity rejection should be mapped");
        assert!(
            matches!(err, Some(Error::SlowDown)),
            "pool metadata heal must preserve target reservation admission: {err:?}"
        );
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn pool_meta_heal_keeps_per_target_capacity_admission() {
        let (_temp_dirs, store, _other_store) = test_three_pool_stores_with_isolated_node_contexts(None).await;
        let layout = DecommissionErasureLayout { data: 1, parity: 0 };
        set_decommission_capacity_info_overrides_for_test(
            store.id,
            vec![vec![
                DecommissionPoolCapacityInfo::for_test(0, layout, 0, 30, 30),
                DecommissionPoolCapacityInfo::for_test(1, layout, 60, 60, 0),
                DecommissionPoolCapacityInfo::for_test(2, layout, 60, 60, 0),
            ]],
        );
        store
            .save_current_pool_meta_for_decommission_start(&[0], Vec::new())
            .await
            .expect("pool metadata reservation should activate");
        {
            let pool_meta = store.pool_meta.read().await;
            let reservation = pool_meta.pools[0]
                .decommission
                .as_ref()
                .and_then(|info| info.capacity_reservation.as_ref())
                .expect("pool metadata reservation should be durable");
            assert_eq!(reservation.targets.len(), 1);
            assert_eq!(reservation.targets[0].pool_index, 1);
        }

        let missing_disk = remove_pool_meta_shard(&store, 2).await;

        let (result, err) = tokio::time::timeout(
            std::time::Duration::from_secs(30),
            store.handle_heal_object(RUSTFS_META_BUCKET, POOL_META_NAME, "", &HealOpts::default()),
        )
        .await
        .expect("unscoped pool metadata heal should complete")
        .expect("unscoped pool metadata heal should return a mapped result");

        assert!(err.is_none(), "an admitted target should let unscoped metadata heal succeed: {err:?}");
        assert_eq!(result.object, POOL_META_NAME);
        assert!(
            missing_disk.read_xl(RUSTFS_META_BUCKET, POOL_META_NAME, false).await.is_ok(),
            "unscoped metadata heal should repair the admitted target while another target is reserved"
        );
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn targeted_heal_keeps_target_lock_for_different_lock_domain() {
        let (_temp_dirs, store, other_store) = test_three_pool_stores_with_isolated_node_contexts(None).await;
        let bucket = format!("heal-lock-domain-{}", Uuid::new_v4().simple());
        let object = "different-domain-heal.bin";
        store
            .make_bucket(&bucket, &MakeBucketOptions::default())
            .await
            .expect("heal lock-domain bucket should be created");

        let target_set = other_store.pools[1].get_disks(0);
        let mut reader = PutObjReader::from_vec(b"different lock domain body".to_vec());
        target_set
            .put_object(&bucket, object, &mut reader, &ObjectOptions::default())
            .await
            .expect("heal lock-domain fixture should be written");
        let missing_disk = target_set.disks.read().await[0]
            .clone()
            .expect("heal lock-domain fixture disk should be online");
        missing_disk
            .delete(
                &bucket,
                object,
                DeleteOptions {
                    recursive: true,
                    immediate: true,
                    ..Default::default()
                },
            )
            .await
            .expect("heal lock-domain fixture shard should be removed");
        assert!(
            missing_disk.read_xl(&bucket, object, false).await.is_err(),
            "heal lock-domain fixture must start with one missing metadata copy"
        );

        let layout = DecommissionErasureLayout { data: 1, parity: 0 };
        set_decommission_capacity_info_overrides_for_test(
            store.id,
            vec![vec![
                DecommissionPoolCapacityInfo::for_test(0, layout, 0, 30, 30),
                DecommissionPoolCapacityInfo::for_test(1, layout, 0, 100, 100),
                DecommissionPoolCapacityInfo::for_test(2, layout, 60, 60, 0),
            ]],
        );
        store
            .save_current_pool_meta_for_decommission_start(&[0], Vec::new())
            .await
            .expect("heal lock-domain reservation should activate");
        {
            let pool_meta = store.pool_meta.read().await;
            let reservation = pool_meta.pools[0]
                .decommission
                .as_ref()
                .and_then(|info| info.capacity_reservation.as_ref())
                .expect("heal lock-domain reservation should be durable");
            assert_eq!(reservation.targets.len(), 1);
            assert_eq!(reservation.targets[0].pool_index, 2);
        }
        let pool_meta = store.pool_meta.read().await.clone();
        *other_store.pool_meta.write().await = pool_meta;

        let fixed_set = other_store.pools[0].disk_set[0].clone();
        assert!(
            !fixed_set.shares_namespace_lock_domain(&target_set).await,
            "heal fixture must use different fixed and target lock domains"
        );
        let barrier = DecommissionCapacityLockOrderBarrier::install(store.id, other_store.id);
        let heal_store = Arc::clone(&other_store);
        let heal_bucket = bucket.clone();
        let heal_object = object.to_string();
        let mut heal = tokio::spawn(async move {
            heal_store
                .handle_heal_object(
                    &heal_bucket,
                    &heal_object,
                    "",
                    &HealOpts {
                        pool: Some(1),
                        set: Some(0),
                        ..Default::default()
                    },
                )
                .await
        });

        barrier.wait_until_external_capacity_released().await;
        barrier.wait_until_external_heal_target_lock_attempted().await;

        let (_, err) = tokio::time::timeout(std::time::Duration::from_secs(30), &mut heal)
            .await
            .expect("targeted heal should finish after target lock release")
            .expect("targeted heal task should not panic")
            .expect("targeted heal should complete");
        assert!(err.is_none(), "targeted heal should repair after target lock release: {err:?}");
        assert!(
            missing_disk.read_xl(&bucket, object, false).await.is_ok(),
            "targeted heal should rewrite the missing shard after the target lock is released"
        );
        drop(barrier);
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn targeted_heal_reuses_shared_target_lock_without_reentrant_lock() {
        let (_temp_dirs, store, _other_store) = test_three_pool_stores_with_isolated_node_contexts(None).await;
        let bucket = format!("heal-shared-lock-domain-{}", Uuid::new_v4().simple());
        let object = "shared-domain-heal.bin";
        store
            .make_bucket(&bucket, &MakeBucketOptions::default())
            .await
            .expect("shared heal lock-domain bucket should be created");

        let target_set = store.pools[0].get_disks(0);
        let mut reader = PutObjReader::from_vec(b"shared lock domain body".to_vec());
        target_set
            .put_object(&bucket, object, &mut reader, &ObjectOptions::default())
            .await
            .expect("shared heal lock-domain fixture should be written");
        let missing_disk = target_set.disks.read().await[0]
            .clone()
            .expect("shared heal lock-domain fixture disk should be online");
        missing_disk
            .delete(
                &bucket,
                object,
                DeleteOptions {
                    recursive: true,
                    immediate: true,
                    ..Default::default()
                },
            )
            .await
            .expect("shared heal lock-domain fixture shard should be removed");
        assert!(
            missing_disk.read_xl(&bucket, object, false).await.is_err(),
            "shared heal lock-domain fixture must start with one missing metadata copy"
        );

        let layout = DecommissionErasureLayout { data: 1, parity: 0 };
        set_decommission_capacity_info_overrides_for_test(
            store.id,
            vec![vec![
                DecommissionPoolCapacityInfo::for_test(0, layout, 0, 100, 100),
                DecommissionPoolCapacityInfo::for_test(1, layout, 60, 60, 0),
                DecommissionPoolCapacityInfo::for_test(2, layout, 0, 30, 30),
            ]],
        );
        store
            .save_current_pool_meta_for_decommission_start(&[2], Vec::new())
            .await
            .expect("shared heal lock-domain reservation should activate");
        {
            let pool_meta = store.pool_meta.read().await;
            let reservation = pool_meta.pools[2]
                .decommission
                .as_ref()
                .and_then(|info| info.capacity_reservation.as_ref())
                .expect("shared heal lock-domain reservation should be durable");
            assert_eq!(reservation.targets.len(), 1);
            assert_eq!(reservation.targets[0].pool_index, 1);
            assert_eq!(reservation.targets[0].reserved_physical_bytes, 60);
        }

        let fixed_set = store.pools[0].disk_set[0].clone();
        assert!(
            fixed_set.shares_namespace_lock_domain(&target_set).await,
            "shared heal fixture must use one fixed and target lock domain"
        );

        let barrier = DecommissionCapacityLockOrderBarrier::install(store.id, store.id);
        let heal_store = Arc::clone(&store);
        let heal_bucket = bucket.clone();
        let heal_object = object.to_string();
        let mut heal = tokio::spawn(async move {
            heal_store
                .handle_heal_object(
                    &heal_bucket,
                    &heal_object,
                    "",
                    &HealOpts {
                        pool: Some(0),
                        set: Some(0),
                        ..Default::default()
                    },
                )
                .await
        });

        barrier.wait_until_external_capacity_released().await;
        barrier.wait_until_external_heal_operation_started().await;
        let (_, err) = tokio::time::timeout(std::time::Duration::from_secs(5), &mut heal)
            .await
            .expect("shared-domain heal must not reenter the fixed namespace lock")
            .expect("shared-domain heal task should not panic")
            .expect("shared-domain heal should complete");
        assert!(err.is_none(), "shared-domain heal should repair after admission: {err:?}");
        assert!(
            missing_disk.read_xl(&bucket, object, false).await.is_ok(),
            "shared-domain heal should rewrite the missing shard"
        );
        drop(barrier);
    }

    #[tokio::test]
    async fn scoped_heal_object_defers_when_requested_pool_is_suspended() {
        let mut store = minimal_heal_store().await;
        store.pool_meta = RwLock::new(PoolMeta {
            pools: vec![
                PoolStatus {
                    id: 0,
                    cmd_line: "pool-0".to_string(),
                    last_update: OffsetDateTime::UNIX_EPOCH,
                    decommission: None,
                },
                PoolStatus {
                    id: 1,
                    cmd_line: "pool-1".to_string(),
                    last_update: OffsetDateTime::UNIX_EPOCH,
                    decommission: Some(PoolDecommissionInfo {
                        start_time: Some(OffsetDateTime::UNIX_EPOCH),
                        ..Default::default()
                    }),
                },
            ],
            ..Default::default()
        });

        let (_, err) = store
            .handle_heal_object(
                "bucket",
                "object",
                "",
                &HealOpts {
                    pool: Some(1),
                    set: Some(0),
                    ..Default::default()
                },
            )
            .await
            .expect("suspended pool should return a deferred heal result");

        assert!(matches!(err, Some(StorageError::SlowDown)));

        let (_, err) = store
            .handle_heal_object(
                "bucket",
                "object",
                "",
                &HealOpts {
                    set: Some(1),
                    ..Default::default()
                },
            )
            .await
            .expect("unscoped heal should return the active pool result");

        assert!(matches!(err, Some(StorageError::InvalidArgument(_, ref field, _)) if field == "set"));

        let err = store
            .handle_heal_object(
                "bucket",
                "object",
                "",
                &HealOpts {
                    pool: Some(1),
                    set: Some(1),
                    ..Default::default()
                },
            )
            .await
            .expect_err("invalid set scope should fail before suspended pool deferral");

        assert!(matches!(err, StorageError::InvalidArgument(_, ref field, _) if field == "set"));

        {
            let mut pool_meta = store.pool_meta.write().await;
            let decommission = pool_meta.pools[1]
                .decommission
                .as_mut()
                .expect("test pool should have decommission state");
            decommission.complete = true;
        }
        let (_, err) = store
            .handle_heal_object(
                "bucket",
                "object",
                "",
                &HealOpts {
                    pool: Some(1),
                    set: Some(0),
                    ..Default::default()
                },
            )
            .await
            .expect("completed pool should return a terminal heal result");

        assert!(matches!(
            err,
            Some(StorageError::InvalidArgument(_, ref field, ref reason))
                if field == "pool" && reason.contains("completed decommission")
        ));

        for canceled in [false, true] {
            {
                let mut pool_meta = store.pool_meta.write().await;
                let decommission = pool_meta.pools[1]
                    .decommission
                    .as_mut()
                    .expect("test pool should have decommission state");
                decommission.complete = false;
                decommission.failed = !canceled;
                decommission.canceled = canceled;
            }
            let (_, err) = store
                .handle_heal_object(
                    "bucket",
                    "object",
                    "",
                    &HealOpts {
                        pool: Some(1),
                        set: Some(0),
                        ..Default::default()
                    },
                )
                .await
                .expect("clearable terminal pool should return a deferred heal result");

            assert!(matches!(err, Some(StorageError::SlowDown)));
        }
    }

    #[tokio::test]
    async fn scoped_heal_bucket_blocks_before_dispatch_when_pool_is_suspended() {
        let mut store = minimal_heal_store().await;
        store.pool_meta = RwLock::new(PoolMeta {
            pools: vec![
                PoolStatus {
                    id: 0,
                    cmd_line: "pool-0".to_string(),
                    last_update: OffsetDateTime::UNIX_EPOCH,
                    decommission: None,
                },
                PoolStatus {
                    id: 1,
                    cmd_line: "pool-1".to_string(),
                    last_update: OffsetDateTime::UNIX_EPOCH,
                    decommission: Some(PoolDecommissionInfo {
                        start_time: Some(OffsetDateTime::UNIX_EPOCH),
                        ..Default::default()
                    }),
                },
            ],
            ..Default::default()
        });

        let err = store
            .handle_heal_bucket(
                "bucket",
                &HealOpts {
                    pool: Some(1),
                    ..Default::default()
                },
            )
            .await
            .expect_err("suspended pool must be blocked before bucket-heal fan-out");
        assert_eq!(err, Error::SlowDown);

        store.pool_meta.write().await.pools[1]
            .decommission
            .as_mut()
            .expect("decommission state should exist")
            .complete = true;
        let err = store
            .handle_heal_bucket(
                "bucket",
                &HealOpts {
                    pool: Some(1),
                    ..Default::default()
                },
            )
            .await
            .expect_err("completed pool must remain fenced from bucket heal");
        assert!(
            matches!(err, StorageError::InvalidArgument(_, ref field, ref reason)
                if field == "pool" && reason.contains("completed decommission")),
            "unexpected completed-pool error: {err:?}"
        );
    }

    #[tokio::test]
    async fn bucket_heal_blocks_before_dispatch_after_unreadable_pool_meta_replica() {
        let store = minimal_heal_store().await;
        store.pool_meta_save_gate.lock().await.observe_replicas(PoolMetaReplicaState {
            needs_repair: true,
            repair_write_safe: false,
        });

        let err = store
            .handle_heal_bucket("bucket", &HealOpts::default())
            .await
            .expect_err("bucket heal must stay blocked until restart after an unreadable replica");

        assert!(
            err.to_string()
                .contains("restart after all replicas are readable and consistent")
        );
    }

    #[tokio::test]
    async fn bucket_heal_releases_save_gate_before_peer_dispatch_and_holds_movement_snapshot() {
        let mut store = minimal_heal_store().await;
        let started = Arc::new(tokio::sync::Notify::new());
        let release = Arc::new(tokio::sync::Notify::new());
        let peer: Box<dyn PeerS3Client> = Box::new(BlockingHealPeer {
            started: started.clone(),
            release: release.clone(),
        });
        store.peer_sys.clients = vec![Arc::new(peer)];
        let store = Arc::new(store);
        let movement_gate = store.ctx.data_movement_operation_gate();
        let mut heal = tokio::spawn({
            let store = store.clone();
            async move { store.handle_heal_bucket("bucket", &HealOpts::default()).await }
        });

        tokio::time::timeout(std::time::Duration::from_secs(1), started.notified())
            .await
            .expect("peer dispatch should start");
        assert!(
            store.pool_meta_save_gate.try_lock().is_ok(),
            "coordinator must release its local save gate before waiting for peers"
        );
        assert!(
            movement_gate.try_write().is_err(),
            "bucket heal must hold the movement snapshot through peer dispatch"
        );

        release.notify_one();
        tokio::time::timeout(std::time::Duration::from_secs(1), &mut heal)
            .await
            .expect("bucket heal should finish after peer release")
            .expect("bucket heal task should not panic")
            .expect("bucket heal should succeed");
    }

    #[tokio::test]
    async fn bucket_heal_local_fanout_does_not_reenter_movement_read_behind_queued_writer() {
        let mut store = minimal_heal_store().await;
        let movement_gate = store.ctx.data_movement_operation_gate();
        let writer_queued = Arc::new(tokio::sync::Notify::new());
        let writer_acquired = Arc::new(tokio::sync::Notify::new());
        let peer: Box<dyn PeerS3Client> = Box::new(WriterQueuedLocalHealPeer {
            movement_gate: movement_gate.clone(),
            writer_queued: writer_queued.clone(),
            writer_acquired: writer_acquired.clone(),
        });
        store.peer_sys.clients = vec![Arc::new(peer)];
        let store = Arc::new(store);
        let mut heal = tokio::spawn({
            let store = store.clone();
            async move { store.handle_heal_bucket("bucket", &HealOpts::default()).await }
        });

        tokio::time::timeout(std::time::Duration::from_secs(1), writer_queued.notified())
            .await
            .expect("movement writer should queue during local peer lookup");
        tokio::time::timeout(std::time::Duration::from_secs(1), &mut heal)
            .await
            .expect("local fan-out must not reenter movement read behind the queued writer")
            .expect("bucket heal task should not panic")
            .expect("bucket heal should succeed");
        tokio::time::timeout(std::time::Duration::from_secs(1), writer_acquired.notified())
            .await
            .expect("queued movement writer should proceed after bucket heal releases its read guard");
    }

    #[tokio::test]
    #[serial_test::serial]
    async fn unscoped_heal_object_suspended_owner_semantics() {
        let (_temp_dir, store, shutdown) = multi_pool_heal_store().await;
        let bucket = format!("heal-owner-{}", Uuid::new_v4().simple());
        let active_object = "active-owner";
        let suspended_only_object = "suspended-only";
        let duplicate_object = "duplicate-owner";
        let marker_object = "marker-owner";
        let quorum_object = "quorum-owner";
        store
            .make_bucket(&bucket, &MakeBucketOptions::default())
            .await
            .expect("bucket should be created in all pools");

        let mut active_reader = PutObjReader::from_vec(b"active owner".to_vec());
        store.pools[0]
            .put_object(
                &bucket,
                active_object,
                &mut active_reader,
                &ObjectOptions {
                    // Drain the rename tail before removing a physical shard.
                    write_completion: crate::object_api::WriteCompletion::TailDrained,
                    ..Default::default()
                },
            )
            .await
            .expect("active owner object should be written");
        let active_disks = store.pools[0].disk_set[0].disks.read().await.clone();
        let missing_active_disk = active_disks[0].clone().expect("active disk should be online");
        missing_active_disk
            .delete(
                &bucket,
                active_object,
                DeleteOptions {
                    recursive: true,
                    immediate: true,
                    ..Default::default()
                },
            )
            .await
            .expect("active owner shard should be removed for repair");
        assert!(
            missing_active_disk.read_xl(&bucket, active_object, false).await.is_err(),
            "the active owner fixture must start with one missing metadata copy"
        );

        let mut suspended_reader = PutObjReader::from_vec(b"suspended owner".to_vec());
        store.pools[1]
            .put_object(&bucket, suspended_only_object, &mut suspended_reader, &ObjectOptions::default())
            .await
            .expect("suspended owner object should be written");
        for (pool_index, mod_time) in [1_i64, 2_i64].into_iter().enumerate() {
            let mut duplicate_reader = PutObjReader::from_vec(format!("duplicate-pool-{pool_index}").into_bytes());
            store.pools[pool_index]
                .put_object(
                    &bucket,
                    duplicate_object,
                    &mut duplicate_reader,
                    &ObjectOptions {
                        mod_time: Some(OffsetDateTime::UNIX_EPOCH + time::Duration::seconds(mod_time)),
                        write_completion: crate::object_api::WriteCompletion::TailDrained,
                        ..Default::default()
                    },
                )
                .await
                .expect("duplicate owner object should be written");
        }
        let duplicate_missing_disk = store.pools[0].disk_set[0].disks.read().await[0]
            .clone()
            .expect("duplicate active owner disk should be online");
        duplicate_missing_disk
            .delete(
                &bucket,
                duplicate_object,
                DeleteOptions {
                    recursive: true,
                    immediate: true,
                    ..Default::default()
                },
            )
            .await
            .expect("duplicate active owner shard should be removed for repair");
        let history_version = Uuid::new_v4();
        let mut history_reader = PutObjReader::from_vec(b"marker history".to_vec());
        store.pools[0]
            .put_object(
                &bucket,
                marker_object,
                &mut history_reader,
                &ObjectOptions {
                    versioned: true,
                    version_id: Some(history_version.to_string()),
                    mod_time: Some(OffsetDateTime::UNIX_EPOCH + time::Duration::seconds(1)),
                    ..Default::default()
                },
            )
            .await
            .expect("versioned marker history should be written");
        store.pools[0]
            .delete_object(
                &bucket,
                marker_object,
                ObjectOptions {
                    versioned: true,
                    mod_time: Some(OffsetDateTime::UNIX_EPOCH + time::Duration::seconds(2)),
                    ..Default::default()
                },
            )
            .await
            .expect("delete marker should be written");
        let mut quorum_reader = PutObjReader::from_vec(b"quorum boundary".to_vec());
        store.pools[0]
            .put_object(&bucket, quorum_object, &mut quorum_reader, &ObjectOptions::default())
            .await
            .expect("quorum boundary object should be written");
        {
            let mut pool_meta = store.pool_meta.write().await;
            let mut next = PoolMeta::new(&store.pools, &pool_meta);
            next.pools[1].decommission = Some(PoolDecommissionInfo {
                start_time: Some(OffsetDateTime::UNIX_EPOCH),
                ..Default::default()
            });
            *pool_meta = next;
        }

        let (_, duplicate_owner) = store
            .get_latest_object_info_with_idx(&bucket, duplicate_object, &ObjectOptions::default())
            .await
            .expect("duplicate owner should resolve");
        assert_eq!(duplicate_owner, 1, "latest duplicate must win when all pools are eligible");
        let (_, active_duplicate_owner) = store
            .get_latest_object_info_with_idx(
                &bucket,
                duplicate_object,
                &ObjectOptions {
                    skip_decommissioned: true,
                    ..Default::default()
                },
            )
            .await
            .expect("active duplicate owner should resolve");
        assert_eq!(
            active_duplicate_owner, 0,
            "suspended duplicate must be excluded from active owner selection"
        );
        let (duplicate_result, duplicate_err) = store
            .handle_heal_object(&bucket, duplicate_object, "", &HealOpts::default())
            .await
            .expect("duplicate owner heal should complete through the production path");
        assert_eq!(duplicate_result.object, duplicate_object);
        assert!(duplicate_err.is_none(), "active duplicate should be repaired: {duplicate_err:?}");
        assert!(
            duplicate_missing_disk.read_xl(&bucket, duplicate_object, false).await.is_ok(),
            "production heal must repair the active duplicate owner rather than the suspended owner"
        );
        let (marker_info, marker_owner) = store
            .get_latest_object_info_with_idx(
                &bucket,
                marker_object,
                &ObjectOptions {
                    skip_decommissioned: true,
                    versioned: true,
                    ..Default::default()
                },
            )
            .await
            .expect("latest delete marker should resolve");
        assert_eq!(marker_owner, 0);
        assert!(marker_info.delete_marker, "latest version must preserve delete-marker semantics");

        let (active_result, active_err) = store
            .handle_heal_object(&bucket, active_object, "", &HealOpts::default())
            .await
            .expect("unscoped active-owner heal should complete");
        assert_eq!(active_result.object, active_object);
        assert!(active_err.is_none(), "active owner must be selected even with a suspended pool");
        assert!(
            missing_active_disk.read_xl(&bucket, active_object, false).await.is_ok(),
            "active owner heal must write the missing disk metadata: result={active_result:?}, err={active_err:?}"
        );
        assert!(
            store.pools[1]
                .get_object_info(&bucket, active_object, &ObjectOptions::default())
                .await
                .is_err(),
            "the suspended pool must not be written for an active-owner object"
        );

        let (suspended_result, suspended_err) = store
            .handle_heal_object(&bucket, suspended_only_object, "", &HealOpts::default())
            .await
            .expect("unscoped suspended-only heal should return a terminal result");
        assert!(suspended_result.object.is_empty());
        assert!(matches!(suspended_err, Some(Error::FileNotFound)));
        assert!(
            store.pools[1]
                .get_object_info(&bucket, suspended_only_object, &ObjectOptions::default())
                .await
                .is_ok(),
            "suspended-only data must remain untouched when unscoped heal reports absent"
        );

        let (_, explicit_err) = store
            .handle_heal_object(
                &bucket,
                suspended_only_object,
                "",
                &HealOpts {
                    pool: Some(1),
                    ..Default::default()
                },
            )
            .await
            .expect("explicit suspended-owner heal should return a mapped error");
        assert!(matches!(explicit_err, Some(Error::SlowDown)));

        let original_quorum_disks = store.pools[0].disk_set[0].disks.read().await.clone();
        let surviving_quorum_disk = original_quorum_disks[3].clone();
        *store.pools[0].disk_set[0].disks.write().await = vec![None, None, None, surviving_quorum_disk];
        let (_, quorum_err) = store
            .handle_heal_object(&bucket, quorum_object, "", &HealOpts::default())
            .await
            .expect("quorum boundary heal should return a mapped result");
        *store.pools[0].disk_set[0].disks.write().await = original_quorum_disks;
        let quorum_err = quorum_err
            .as_ref()
            .expect("heal must fail closed when capacity admission cannot verify pool metadata");
        let quorum_failure = quorum_err
            .pool_metadata_failure()
            .expect("capacity admission failure should preserve typed pool metadata context");
        assert_eq!(
            quorum_failure.kind,
            crate::error::PoolMetadataFailure::ReadUnavailable,
            "read-only capacity admission failure must remain retryable"
        );
        assert_eq!(quorum_failure.operation, "target capacity admission failed");
        assert_eq!(quorum_failure.phase, "pool_read");
        assert!(
            store.pool_meta_writes_ready().await,
            "read-only capacity admission failure must not latch the pool metadata writer"
        );
        shutdown.cancel();
    }

    #[tokio::test]
    async fn handle_heal_format_continues_after_a_pool_error() {
        let canonical_format = FormatV3::new(1, 3);
        let mut foreign_format = canonical_format.clone();
        foreign_format.id = Uuid::new_v4();
        let mut temp_dirs = Vec::new();
        let mut endpoints = Vec::new();
        let mut disks = Vec::new();

        for disk_index in 0..3 {
            let temp_dir = tempfile::tempdir().expect("temporary disk root should be created");
            let mut endpoint = Endpoint::try_from(temp_dir.path().to_str().expect("temporary path should be UTF-8"))
                .expect("temporary endpoint should parse");
            endpoint.set_pool_index(0);
            endpoint.set_set_index(0);
            endpoint.set_disk_index(disk_index);
            let disk = new_disk(
                &endpoint,
                &DiskOption {
                    cleanup: false,
                    health_check: false,
                },
            )
            .await
            .expect("temporary disk should open");
            let mut disk_format = foreign_format.clone();
            disk_format.erasure.this = foreign_format.erasure.sets[0][disk_index];
            save_format_file(&Some(disk.clone()), &Some(disk_format))
                .await
                .expect("foreign format should be written");
            temp_dirs.push(temp_dir);
            endpoints.push(endpoint);
            disks.push(Some(disk));
        }

        let pool_endpoints = PoolEndpoints {
            legacy: false,
            set_count: 1,
            drives_per_set: 3,
            endpoints: Endpoints::from(endpoints),
            cmd_line: "foreign-format-majority-test".to_string(),
            platform: "test".to_string(),
        };
        let pool = Sets::new(disks, &pool_endpoints, &canonical_format, 0, 1)
            .await
            .expect("test pool should build around the cached canonical format");

        let mut recoverable_format = FormatV3::new(1, 3);
        recoverable_format.id = canonical_format.id;
        let mut recoverable_temp_dirs = Vec::new();
        let mut recoverable_endpoints = Vec::new();
        let mut recoverable_disks = Vec::new();
        let mut unformatted_disk = None;
        for disk_index in 0..3 {
            let temp_dir = tempfile::tempdir().expect("temporary disk root should be created");
            let mut endpoint = Endpoint::try_from(temp_dir.path().to_str().expect("temporary path should be UTF-8"))
                .expect("temporary endpoint should parse");
            endpoint.set_pool_index(1);
            endpoint.set_set_index(0);
            endpoint.set_disk_index(disk_index);
            let disk = new_disk(
                &endpoint,
                &DiskOption {
                    cleanup: false,
                    health_check: false,
                },
            )
            .await
            .expect("temporary disk should open");
            if disk_index < 2 {
                let mut disk_format = recoverable_format.clone();
                disk_format.erasure.this = recoverable_format.erasure.sets[0][disk_index];
                save_format_file(&Some(disk.clone()), &Some(disk_format))
                    .await
                    .expect("recoverable format should be written");
            } else {
                unformatted_disk = Some(disk.clone());
            }
            recoverable_temp_dirs.push(temp_dir);
            recoverable_endpoints.push(endpoint);
            recoverable_disks.push(Some(disk));
        }
        let recoverable_pool_endpoints = PoolEndpoints {
            legacy: false,
            set_count: 1,
            drives_per_set: 3,
            endpoints: Endpoints::from(recoverable_endpoints),
            cmd_line: "recoverable-format-test".to_string(),
            platform: "test".to_string(),
        };
        let recoverable_pool = Sets::new(recoverable_disks, &recoverable_pool_endpoints, &recoverable_format, 1, 1)
            .await
            .expect("recoverable test pool should build");

        let endpoint_pools = EndpointServerPools::from(vec![pool_endpoints.clone(), recoverable_pool_endpoints.clone()]);
        let store = ECStore {
            id: canonical_format.id,
            disk_map: HashMap::new(),
            pools: vec![pool, recoverable_pool],
            peer_sys: S3PeerSys::new(&endpoint_pools),
            pool_meta: RwLock::new(PoolMeta::default()),
            rebalance_meta: RwLock::new(None),
            decommission_cancelers: RwLock::new(Vec::new()),
            start_gate: Mutex::new(()),
            pool_meta_save_gate: Mutex::default(),
            ctx: crate::runtime::instance::bootstrap_ctx(),
            bucket_fence_registry: std::sync::Arc::default(),
        };

        let err = store
            .handle_heal_format(false)
            .await
            .expect_err("missing pool metadata must fail closed before format writes");
        assert!(err.to_string().contains("no durable bootstrap identity or pool.bin replica"));
        store
            .ensure_pool_meta_side_effects_safe("missing format-heal metadata side effect")
            .await
            .expect_err("missing metadata must latch the format-heal write gate");

        let pool_meta = PoolMeta::new(&store.pools, &PoolMeta::default());
        pool_meta
            .save_for_startup(store.pools.clone())
            .await
            .expect("pool metadata should be persisted before format heal");
        *store.pool_meta_save_gate.lock().await = PoolMetaWriteState::default();

        let (result, err) = store
            .handle_heal_format(false)
            .await
            .expect("format heal should return the typed pool error");
        assert!(
            matches!(err, Some(StorageError::CorruptedFormat)),
            "foreign format majority must not be downgraded to a successful heal: {err:?}"
        );
        assert_eq!(result.disk_count, 3, "the recoverable pool should still be inspected");
        let healed = load_format_erasure(&unformatted_disk.expect("the unformatted disk handle should be retained"), true)
            .await
            .expect("the later pool should be healed despite the first pool error");
        assert_eq!(healed.erasure.this, recoverable_format.erasure.sets[0][2]);

        let mut completed_meta = PoolMeta::new(&store.pools, &PoolMeta::default());
        for status in &mut completed_meta.pools {
            status.decommission = Some(PoolDecommissionInfo {
                complete: true,
                ..Default::default()
            });
        }
        completed_meta
            .save(store.pools.clone())
            .await
            .expect("completed pool metadata should be persisted");
        let (_, err) = store
            .handle_heal_format(false)
            .await
            .expect("completed pools should be reported as a no-op");
        assert!(matches!(err, Some(StorageError::NoHealRequired)));
    }
}