| = Disk Compression |
| |
| Disk compression refers to the process of compressing data pages when they are written to disk, reducing the size of the on-disk storage. |
| The pages are kept in memory uncompressed, but when the data is flushed to disk it is compressed using the configured algorithm. |
| This applies only to data pages that are stored to the persistent storage and does not compress indexes or WAL records. |
| link:[WAL records compression] can be enabled separately. |
| |
| Disk page compression can be enabled on a per cache basis in the cache configuration. |
| The cache must reside in a persistent link:[data region]. |
| There is no option to enable disk page compression globally at the moment. |
| Moreover, the following prerequisites must be met: |
| |
| * Set the `pageSize` property in your data storage configuration to at least 2 times the page size of your file system. It means that the page size must be either 8K or 16K. |
| * Enable the `ignite-compress` module. |
| |
| To enable disk page compression for a cache, provide one of the available compression algorithms in the cache configuration, as shown in the following example: |
| |
| |
| [tabs] |
| -- |
| tab:XML[] |
| |
| [source, xml] |
| ---- |
| include::code-snippets/xml/disk-compression.xml[tags=ignite-config;!discovery, indent=0] |
| ---- |
| |
| tab:Java[] |
| |
| [source, java] |
| ---- |
| include::{javaCodeDir}/DiskCompression.java[tags=configuration, indent=0] |
| ---- |
| |
| tab:C#/.NET[] |
| |
| tab:C++[unsupported] |
| |
| -- |
| |
| == Supported Algorithms |
| |
| The supported compression algorithms include: |
| |
| * `ZSTD` — supports compression levels from -131072 to 22 (default: 3). |
| * `LZ4` — supports compression levels from 0 to 17 (default: 0). |
| * `SNAPPY` — the Snappy algorithm. |
| * `SKIP_GARBAGE` — this algorithm only extracts useful data from half-filled pages and does not compress the data. |
