Summary:
After releasing a snapshot, it checks whether it is suitable to trigger bottom compactions.
When disabling auto compactions, it may still schedule compaction when releasing a snapshot. Whereas no compaction job will be actually handled, so the state of LSM is not changed and compaction will be triggered again and again every time releasing a snapshot.
Too frequent compactions lead to high CPU usage and high db_mutex lock contention which affects foreground write duration finally.
Pull Request resolved: https://github.com/facebook/rocksdb/pull/7267
Test Plan:
- make check
- manual test
Reviewed By: akankshamahajan15
Differential Revision: D23252880
Pulled By: ajkr
fbshipit-source-id: 4431e071a35d9912a2a3592875db27bae521434b
main
Connor19964 years agocommitted byFacebook GitHub Bot
* Fix a bug where a level refitting in CompactRange() might race with an automatic compaction that puts the data to the target level of the refitting. The bug has been there for years.
* Fix a bug where a level refitting in CompactRange() might race with an automatic compaction that puts the data to the target level of the refitting. The bug has been there for years.
* BackupEngine::CreateNewBackup could fail intermittently with non-OK status when backing up a read-write DB configured with a DBOptions::file_checksum_gen_factory. This issue has been worked-around such that CreateNewBackup should succeed, but (until fully fixed) BackupEngine might not see all checksums available in the DB.
* BackupEngine::CreateNewBackup could fail intermittently with non-OK status when backing up a read-write DB configured with a DBOptions::file_checksum_gen_factory. This issue has been worked-around such that CreateNewBackup should succeed, but (until fully fixed) BackupEngine might not see all checksums available in the DB.
* Fix a bug where immutable flushed memtable is never destroyed because a memtable is not added to delete list because of refernce hold by super version and super version doesn't switch because of empty delete list. So memory usage increases beyond write_buffer_size + max_write_buffer_size_to_maintain.
* Fix a bug where immutable flushed memtable is never destroyed because a memtable is not added to delete list because of refernce hold by super version and super version doesn't switch because of empty delete list. So memory usage increases beyond write_buffer_size + max_write_buffer_size_to_maintain.
* Fix useless no-op compactions scheduled upon snapshot release when options.disable-auto-compactions = true.
### New Features
### New Features
* A new option `std::shared_ptr<FileChecksumGenFactory> file_checksum_gen_factory` is added to `BackupableDBOptions`. The default value for this option is `nullptr`. If this option is null, the default backup engine checksum function (crc32c) will be used for creating, verifying, or restoring backups. If it is not null and is set to the DB custom checksum factory, the custom checksum function used in DB will also be used for creating, verifying, or restoring backups, in addition to the default checksum function (crc32c). If it is not null and is set to a custom checksum factory different than the DB custom checksum factory (which may be null), BackupEngine will return `Status::InvalidArgument()`.
* A new option `std::shared_ptr<FileChecksumGenFactory> file_checksum_gen_factory` is added to `BackupableDBOptions`. The default value for this option is `nullptr`. If this option is null, the default backup engine checksum function (crc32c) will be used for creating, verifying, or restoring backups. If it is not null and is set to the DB custom checksum factory, the custom checksum function used in DB will also be used for creating, verifying, or restoring backups, in addition to the default checksum function (crc32c). If it is not null and is set to a custom checksum factory different than the DB custom checksum factory (which may be null), BackupEngine will return `Status::InvalidArgument()`.