Commit graph

7 commits

Author SHA1 Message Date
Andrew Kryczka 05c3b8ecac Prepare for specialized interface for row cache (#11620)
Summary:
An internal user wants to implement a key-aware row cache policy. For that, they need to know the components of the cache key, especially the user key component. With a specialized `RowCache` interface, we will be able to tell them the components so they won't have to make assumptions about our internal key schema.

This PR prepares for the specialized `RowCache` interface by updating the migration plan of https://github.com/facebook/rocksdb/issues/11450. I added a release note for the removed APIs and didn't mention the added ones for now.

Pull Request resolved: https://github.com/facebook/rocksdb/pull/11620

Reviewed By: pdillinger

Differential Revision: D47536962

Pulled By: ajkr

fbshipit-source-id: bbee0fc4ad67fc699a66b8f2b4ea4544dd003691
2023-07-18 19:12:58 -07:00
akankshamahajan 94c247bff8 Update HISTORY.md for branch cut for 8.4.fb (#11565)
Summary: Pull Request resolved: https://github.com/facebook/rocksdb/pull/11565

Reviewed By: jowlyzhang, cbi42

Differential Revision: D47027788

Pulled By: akankshamahajan15

fbshipit-source-id: e5e8db2eb21f8aa68fe072f0e1b63b83ba7beb9f
2023-06-26 13:26:15 -07:00
akankshamahajan fbd2f563bb Add an interface to provide support for underlying FS to pass their own buffer during reads (#11324)
Summary:
1. Public API change: Replace `use_async_io`  API in file_system with `SupportedOps` API which is used by underlying FileSystem to indicate to upper layers whether the FileSystem supports different operations introduced in `enum FSSupportedOps `. Right now operations are `async_io` and whether FS will provide its own buffer during reads or not. The api is changed to extend it to various FileSystem operations in one API rather than creating a separate API for each operation.

2. Provide support for underlying FS to pass their own buffer during Reads (async and sync read) instead of using RocksDB provided `scratch` (buffer) in `FSReadRequest`. Currently only MultiRead supports it and later will be extended to other reads as well (point lookup, scan etc). More details in how to enable in file_system.h

Pull Request resolved: https://github.com/facebook/rocksdb/pull/11324

Test Plan: Tested locally

Reviewed By: anand1976

Differential Revision: D44465322

Pulled By: akankshamahajan15

fbshipit-source-id: 9ec9e08f839b5cc815e75d5dade6cd549998d0ec
2023-06-23 11:48:49 -07:00
Alexandre Lavigne 2926e0718c Add missing parameter in C API (#11542)
Summary:
The class `NewCompactOnDeletionCollectorFactory` exposes the parameter `delete_ratio`.

The C API `rocksdb_options_add_compact_on_deletion_collector_factory` does not allow a user to pass a delete ration to be passed down the the C++ class bellow.

The class has default value for the delete ratio which makes it pass the compilation and the tests.

closes https://github.com/facebook/rocksdb/issues/11541

Pull Request resolved: https://github.com/facebook/rocksdb/pull/11542

Reviewed By: ajkr

Differential Revision: D46770908

Pulled By: cbi42

fbshipit-source-id: 7b5162fe459896052e392e2d85a8f6c01db3b464
2023-06-20 13:18:04 -07:00
mayue.fight fa878a0107 Support to create a CF by importing multiple non-overlapping CFs (#11378)
Summary:
The original Feature Request is from [https://github.com/facebook/rocksdb/issues/11317](https://github.com/facebook/rocksdb/issues/11317).
Flink uses rocksdb as the state backend,  all DB options are the same, and the keys of each DB instance are adjacent and there is no key overlap between two db instances.
In the Flink rescaling scenario, it is necessary to quickly split the DB according to a certain key range or quickly merge multiple DBs into one.

This PR is mainly used to quickly merge multiple DBs into one.

We hope to extend the function of `CreateColumnFamilyWithImports` to support creating ColumnFamily by importing multiple ColumnFamily with no overlapping keys.

The import logic is almost the same as `CreateColumnFamilyWithImport`, but it will check whether there is key overlap between CF when importing. The import will fail if there are key overlaps.

Pull Request resolved: https://github.com/facebook/rocksdb/pull/11378

Reviewed By: ajkr

Differential Revision: D46413709

Pulled By: cbi42

fbshipit-source-id: 846d0049fad11c59cf460fa846c345b26c658dfb
2023-06-15 12:25:04 -07:00
Ignat Loskutov 7c67aee4a0 statistics.cc: fix mistype (#11509)
Summary:
Add new tickers: `rocksdb.error.handler.bg.error.count`, `rocksdb.error.handler.bg.io.error.count`, `rocksdb.error.handler.bg.retryable.io.error.count` to replace the misspelled ones: `rocksdb.error.handler.bg.errro.count`, `rocksdb.error.handler.bg.io.errro.count`, `rocksdb.error.handler.bg.retryable.io.errro.count` ('error' instead of 'errro'). Users should switch to use the new tickers before 9.0 release as the misspelled old tickers will be completely removed then.

Pull Request resolved: https://github.com/facebook/rocksdb/pull/11509

Reviewed By: ltamasi

Differential Revision: D46542809

Pulled By: jowlyzhang

fbshipit-source-id: a2a6d8354af46a060de81d40ef6f5336a80bd32e
2023-06-09 13:25:57 -07:00
Peter Dillinger 8848ec92dd Better management of unreleased HISTORY (#11481)
Summary:
See new NOTE in HISTORY.md and unreleased_history/README.txt

Pull Request resolved: https://github.com/facebook/rocksdb/pull/11481

Test Plan: some manual testing on my CentOS 8 system

Reviewed By: jaykorean

Differential Revision: D46233342

Pulled By: pdillinger

fbshipit-source-id: daf59cf3dc907f450b469090dcc481a30a7d7c0d
2023-05-30 16:42:49 -07:00