Struct photondb::PageStoreOptions

#[non_exhaustive]
pub struct PageStoreOptions {

    pub write_buffer_capacity: u32,
    pub max_write_buffers: usize,
    pub use_direct_io: bool,
    pub max_space_amplification_percent: usize,
    pub space_used_high: u64,
    pub file_base_size: usize,
    pub cache_capacity: usize,
    pub cache_estimated_entry_charge: usize,
    pub cache_file_reader_capacity: u64,
    pub prepopulate_cache_on_flush: bool,
    pub separate_hot_cold_files: bool,
    pub compression_on_flush: Compression,
    pub compression_on_cold_compact: Compression,
    pub page_checksum_type: ChecksumType,
}

Options to configure a page store.

Fields (Non-exhaustive)

Non-exhaustive structs could have additional fields added in future. Therefore, non-exhaustive structs cannot be constructed in external crates using the traditional Struct { .. } syntax; cannot be matched against without a wildcard ..; and struct update syntax will not work.

write_buffer_capacity: u32

The capacity of the write buffer. It should be power of two.

Default: 128MB

max_write_buffers: usize

The maxmum number of write buffers.

If there exists too many write buffers, writing will be stalled until at leaset one write buffer is flushed.

Default: 8

use_direct_io: bool

If true, use O_DIRECT to read/write page files.

Default: false

max_space_amplification_percent: usize

The max percentage of the space amplification.

The space amplification is defined as the amount (in percentage) of additional storage needed to store a single byte of data in the database.

Default: 100

space_used_high: u64

The high watermark of the used storage space of the database.

Default: u64::MAX

file_base_size: usize

Target file size for compaction.

Default: 64MB

cache_capacity: usize

The capacity of the page read cache in bytes.

Default: 8 Mib

cache_estimated_entry_charge: usize

The estimated average charge associated with cache entries.

Default: 8 Kib

This is a critical configuration parameter for good performance for page read cache, because having a table size that is fixed at creation time greatly reduces the required synchronization between threads.

• If the estimate is substantially too low (e.g. less than half the true average) then metadata space overhead with be substantially higher (e.g. 200 bytes per entry rather than 100). This can slightly reduce cache hit rates, and slightly reduce access times due to the larger working memory size.
• If the estimate is substantially too high (e.g. 25% higher than the true average) then there might not be sufficient slots in the hash table for both efficient operation and capacity utilization (hit rate). The cache will evict entries to prevent load factors that could dramatically affect lookup times, instead letting the hit rate suffer by not utilizing the full capacity.

cache_file_reader_capacity: u64

The capacity of file_reader cache.

Default: 5000 file_readers.

prepopulate_cache_on_flush: bool

Insert warm pages into PageCache during flush if true.

Default: false

separate_hot_cold_files: bool

Separate page files into hot/cold parts.

This is temproray options for validating.

Default: false

compression_on_flush: Compression

Compression method during flush new file. include hot rewrite.

Default: Snappy.

compression_on_cold_compact: Compression

Compression method during compact cold file.

Default: Zstd(Level3).

page_checksum_type: ChecksumType

ChecksumType for each page.

Default: NONE.

Trait Implementations

impl Clone for Options

fn clone(&self) -> Options

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for Options

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for Options

fn default() -> Self

Returns the “default value” for a type.