[
https://issues.apache.org/jira/browse/SOLR-17153?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
David Smiley updated SOLR-17153:
--------------------------------
Description:
🤖 When a Solr node receives a request for a collection missing from its local
cluster state, it now refreshes aliases and force-updates that collection from
ZooKeeper before failing. Previously only V2 API requests did this; now V1
{{/solr/<collection>/...}} requests do too. This helps when a node's ZK view
lags, e.g. right after collection creation on a heavily loaded cluster.
This issue originally also changed SolrJ's ZkClientClusterStateProvider to
refresh on a miss. That part shipped only in 9.6.0 and was reverted in 9.6.1
(SOLR-17275, performance regression with aliases). The original client-side
goal continues in SOLR-18606, under umbrella SOLR-18154.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
was:
_(Present status: reverted / not accomplished. Â New issues needed to continue)_
Today, CloudSolrClient will locally fail if it's asked to send a request to a
collection that it thinks does not exist due to its local ClusterState view
being out-of-date. Â We shouldn't fail! Â And most SolrCloud tests should then
remove their waitForState calls that follow collection creation! Â Other stale
state matters are out-of-scope.
Proposal: CloudSolrClient shouldn't try and be too smart. Â Always route a
request to Solr (any node); don't presume its state is up-to-date. Â Maybe,
after a response is received, it can check if its state has been updated and if
not then explicitly get a new state. Â Or not if that's too complicated.
Summary: HttpSolrCall: refresh ZK state before failing a request for an
unknown collection (was: CloudSolrClient should not throw "Collection not
found" with an out-dated ClusterState)
> HttpSolrCall: refresh ZK state before failing a request for an unknown
> collection
> ---------------------------------------------------------------------------------
>
> Key: SOLR-17153
> URL: https://issues.apache.org/jira/browse/SOLR-17153
> Project: Solr
> Issue Type: Improvement
> Components: SolrJ
> Reporter: David Smiley
> Priority: Major
> Labels: pull-request-available
> Fix For: 9.6
>
> Time Spent: 6h 20m
> Remaining Estimate: 0h
>
> 🤖 When a Solr node receives a request for a collection missing from its local
> cluster state, it now refreshes aliases and force-updates that collection
> from ZooKeeper before failing. Previously only V2 API requests did this; now
> V1 {{/solr/<collection>/...}} requests do too. This helps when a node's ZK
> view lags, e.g. right after collection creation on a heavily loaded cluster.
> This issue originally also changed SolrJ's ZkClientClusterStateProvider to
> refresh on a miss. That part shipped only in 9.6.0 and was reverted in 9.6.1
> (SOLR-17275, performance regression with aliases). The original client-side
> goal continues in SOLR-18606, under umbrella SOLR-18154.
> 🤖 Generated with [Claude Code](https://claude.com/claude-code)
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
