Make clean up files step configurable for peer-recovery of replicas #92490
Conversation
tlrx
added
>enhancement
:Distributed/Recovery
|
Hi @tlrx, I've created a changelog YAML for you. |
|
Pinging @elastic/es-distributed (Team:Distributed) |
| @@ -495,7 +495,9 @@ public void cleanFiles( | |||
| final Store store = store(); | |||
| store.incRef(); | |||
| try { | |||
| store.cleanupAndVerify("recovery CleanFilesRequestHandler", sourceMetadata); | |||
| if (DiscoveryNode.isStateless(indexShard.indexSettings().getSettings()) == false || indexShard.routingEntry().primary()) { | |||
There was a problem hiding this comment.
I think we need to pass in the node settings, not the index settings to use that method?
There was a problem hiding this comment.
Documentation of IndexSettings#getSettings mentions:
Returns the settings for this index. These settings contain the node and index level settings where
settings that are specified on both index and node level are overwritten by the index settings.
There was a problem hiding this comment.
So I think we're good here: I verified and it does contain the node settings.
There was a problem hiding this comment.
Sure, it will work, but seems more intuitive to use indexShard.indexSettings().getNodeSettings()?
There was a problem hiding this comment.
Oh sorry, I misunderstood your previous comment. Sure, getNodeSettings is more appropriate, thanks for pointing this.
tlrx
added
the
auto-merge
Skip the "clean up and verify" step at the end of files based peer-recovery for replicas.