[
https://issues.apache.org/jira/browse/SOLR-11660?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16259099#comment-16259099
]
Viachaslau Kabak commented on SOLR-11660:
-----------------------------------------
Hi,
sure
1. Backup collection
(http://serverName:SOLR_PORT/solr/admin/collections?action=BACKUP&name=CollectionName&collection=pCollectionName&location=tmpSolrDir)
2. Restore collection (even on another solr Cloud):
2a.
http://serverName:SOLR_PORT/solr/admin/collections?action=DELETE&name=pCollectionName
2b.
http://serverName:SOLR_PORT/solr/admin/collections?action=RESTORE&name=CollectionName&collection=CollectionName&location=tmpSolrDir
3. Try to update index from hybris application
Expected steps: index is updated for leader and for followers
Current results: index is updated only on followers, on the leader there is an
old index, there are warnings in logs:
{noformat}
o.a.s.u.p.DistributedUpdateProcessor Ignoring commit while not ACTIVE - state:
BUFFERING replay: false
{noformat}
Collection reload helps.
Note: if to restart solr on all instances before to update index from hybris
application - new index will be available on all instances and no warnings are
in the log
> Issue while update index in collection after collection restore on Solr Cloud
> -----------------------------------------------------------------------------
>
> Key: SOLR-11660
> URL: https://issues.apache.org/jira/browse/SOLR-11660
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Components: SolrCloud
> Affects Versions: 6.4.1
> Environment: CentOs7.3 Solr 6.4.1 Zookeeper 3.4.6
> Reporter: Viachaslau Kabak
> Priority: Critical
>
> We use to backup and restore one of our collections in solrCloud.
> if not to restore collection but to create it and fill from application - the
> second attempt to renew application index is successful both on leader and
> followers. But is to restore collection and then to renew index from
> application - we have new index on follower and old index on leader.
> The Reload on collection helps to make index up-to-date on leader.
> In logs we have {{o.a.s.u.p.DistributedUpdateProcessor Ignoring commit while
> not ACTIVE - state: BUFFERING replay: false}}
> If to restart solr service before collection update from application - index
> on all solr servers is updated.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]