Deduplicate mappings in persisted cluster state #88479
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
to reflect specifically that there's no action required for an index, specifically
There's no tracking of 'mappings changed' because we're storing things by hash, so mappings are removed and/or added but we can't see 'changes' as a reified thing.
The behavior of writeIncrementalStateAndCommit relies on the previousState that we pass in actually being a correct representation of the on-disk state -- this test did not respect that. Rather, it did clusterState->newClusterState twice in a row, so now we inject a null->clusterState between the two so that everything works right. Also we're doing more writes now, so I bumped the 14 constant to 16.
joegallo
added
WIP
:Distributed Coordination/Cluster Coordination
elasticmachine
added
the
Team:Distributed (Obsolete)
|
Pinging @elastic/es-distributed (Team:Distributed) |
This was referenced Jul 12, 2022
|
I think this is ready for review, @original-brownbear, if you have the time and inclination to take a look. I won't merge it without an explicit ✅ from @DaveCTurner, though, and I know I won't get that until he's back early next week. |
|
Hi @joegallo, I've created a changelog YAML for you. |
DaveCTurner
approved these changes
Jul 18, 2022
| @@ -1629,6 +1894,18 @@ private NodeEnvironment newNodeEnvironment(Path[] dataPaths) throws IOException | |||
| ); | |||
| } | |||
|
|
|||
| private static MappingMetadata randomMappingMetadata(boolean allowNull) { | |||
There was a problem hiding this comment.
nit: slight preference for this being two methods randomMappingMetadata vs randomMappingMetadataOrNull
There was a problem hiding this comment.
👍, I was on the fence about that myself. I'll add a commit.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Replacement for #87126, related to #77466
Following up on #82608, which compacted the network serialization of mappings in the cluster state's metadata. In that PR, mappings are serialized as a map keyed by hash, and then when serializing index metadata the appropriate mapping can be referenced by its hash only, saving network traffic in the (expected common) case of indices with duplicate mappings.
This PR extends that same logic to the on-disk serialization of a cluster state via the PersistedClusterStateService. The efficiency gain in this case is greater, however, because in the case of an index settings change, for example, the mapping and associated hash are unchanged and so the on-disk mapping doesn't need to be re-persisted.
/cc @DaveCTurner