[
https://issues.apache.org/jira/browse/SOLR-12388?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16487208#comment-16487208
]
Shawn Heisey commented on SOLR-12388:
-------------------------------------
Interesting. It's my understanding that SolrCloud goes read-only when ZK
quorum is lost, so it would have to be a particularly unusual network partition
for the described situation to arise. But as noted by the author of Jepsen,
unusual network partitions DO happen in the wild.
> Enable a strict ZooKeeper-connected search request mode, in which search
> requests will fail when the coordinating node can't communicate with ZooKeeper
> -------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: SOLR-12388
> URL: https://issues.apache.org/jira/browse/SOLR-12388
> Project: Solr
> Issue Type: Improvement
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Steve Rowe
> Assignee: Steve Rowe
> Priority: Minor
> Attachments: SOLR-12388.patch
>
>
> Right now, a Solr node will return the results of a search request even if it
> cannot communicate with ZooKeeper at the time it receives the request. This
> may result in stale or incorrect results if there have been major changes to
> the collection structure that the node has not been informed of via
> ZooKeeper. When this happens, as long as all known shards respond, the
> response will succeed, and a {{zkConnected}} header set to {{false}} is
> included in the search response.
> There should be an option to instead fail requests under these conditions, to
> prevent stale or incorrect results.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
