[
https://issues.apache.org/jira/browse/SOLR-8662?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Varun Thacker updated SOLR-8662:
--------------------------------
Attachment: SOLR-8662.patch
Thanks Steve and Noble for the review!
Created the new patch from GIT instead of IntelliJ
In SchemaManager if "updateTimeOutSecs" in 0 or -1 then we retry for 10 minutes
. Earlier it was unlimited. I think capping the unlimited to 10 minutes is fair
enough. Thoughts? It also uses the TimeOut class for the math. Also upped
default timeout to 10 minutes instead of 1 minute earlier
bq. Your changes in ZkSolrResourceLoader.openResource() introduce the
possibility of retrying 10 times to fetch a resource from ZK when it's not there
Good catch! I've fixed that. We won't retry if the path doesn't exist in ZK.
I've kept {{testReloadAndAddSimple()}} for now. In one test run when I was
using AbstractDistribZkTestBase, I'd seen after the reload the session had
expired while reading the managed-schema file. So I had added the test and
retry logic. This was a couple of months back when I had initially worked on
the patch so I don't remember everything.
> SchemaManager doesn't wait correctly for replicas to update
> -----------------------------------------------------------
>
> Key: SOLR-8662
> URL: https://issues.apache.org/jira/browse/SOLR-8662
> Project: Solr
> Issue Type: Bug
> Reporter: Varun Thacker
> Assignee: Varun Thacker
> Attachments: SOLR-8662.patch, SOLR-8662.patch, SOLR-8662.patch
>
>
> Currently in SchemaManager, waitForOtherReplicasToUpdate doesn't execute
> since it gets passed the timeout value instead of the end time
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
