These settings are available in system.settings and are autogenerated from source.
page_cache_block_size
Version history
| Version | Default value | Comment |
|---|---|---|
| 25.5 | 1048576 | Made this setting adjustable on a per-query level. |
Size of file chunks to store in the userspace page cache, in bytes. All reads that go through the cache will be rounded up to a multiple of this size.
This setting can be adjusted on a per-query level basis, but cache entries with different block sizes cannot be reused. Changing this setting effectively invalidates existing entries in the cache.
A higher value, like 1 MiB is good for high-throughput queries, and a lower value, like 64 KiB is good for low-latency point queries.
page_cache_inject_eviction
Version history
| Version | Default value | Comment |
|---|---|---|
| 24.3 | 0 | Added userspace page cache |
Userspace page cache will sometimes invalidate some pages at random. Intended for testing.
page_cache_lookahead_blocks
Version history
| Version | Default value | Comment |
|---|---|---|
| 25.5 | 16 | Made this setting adjustable on a per-query level. |
On userspace page cache miss, read up to this many consecutive blocks at once from the underlying storage, if they’re also not in the cache. Each block is page_cache_block_size bytes.
A higher value is good for high-throughput queries, while low-latency point queries will work better without readahead.
page_cache_max_coalesced_bytes
Version history
| Version | Default value | Comment |
|---|---|---|
| 26.5 | 16777216 | New setting to bound the size of a single coalesced read used to populate the userspace page cache on cache miss. |
When readBigAt populates the userspace page cache, consecutive cache misses are coalesced into a single read from the underlying storage. This setting bounds the size of one coalesced read in bytes; longer miss runs are split into multiple reads. It limits transient memory usage of the temporary buffer under parallel cold reads.
A higher value reduces the number of HTTP requests for cold scans on object storage; a lower value reduces peak transient memory.