[
https://issues.apache.org/jira/browse/SOLR-8633?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15129112#comment-15129112
]
Hoss Man commented on SOLR-8633:
--------------------------------
I suspect that the _intended_ goal here (in processCommit) is to do the same
sort of "force sequential ordering" type logic as doDeleteByQuery -- in which
case we can/should replace the call to {{finish()}} with a call to
{{cmdDistrib.finish()}} ... except that i'm still concerned about
shutdown/closing of clients done by that method, and wonder if _both_
processCommit & processDelete should be calling
{{SolrCmdDistributor.blockAndDoRetries()}} directly? (we would just need to
make that method public)
Hoping folks like [[email protected]] / [~tim.potter] might have some
ideas about this?
> DistributedUpdateProcess.processCommit calls finish() - violates lifecycle,
> causes finish to be called twice (redundent code execution)
> ---------------------------------------------------------------------------------------------------------------------------------------
>
> Key: SOLR-8633
> URL: https://issues.apache.org/jira/browse/SOLR-8633
> Project: Solr
> Issue Type: Bug
> Reporter: Hoss Man
>
> trying to wrap my head around a weird bug in my experiements with SOLR-445, i
> realized that {{DUP.processDelete}} has a direct call to {{finish()}}.
> This violates the normal lifecycle of an UpdateProcessor (finish is only
> suppose to be called exactly once after processing any/all UpdateCommands)
> and could potentially break any UpdateProcessors configured after DUP (or in
> my case: processors configured _before_ DUP that expect to be in charge of
> calling finish, and catching any resulting exceptions, as part of the normal
> life cycle)
> Independent of how it impacts other update processors, this also means that:
> # all the logic in {{DUP.doFinish}} is getting executed twice -- which seems
> kind of expensive/dangerous to me since there is leader initiated recovery
> involved in this method
> # {{SolrCmdDistributor.finish()}} gets called twice, which means
> {{StreamingSolrClients.shutdown()}} gets called twice, which means
> {{ConcurrentUpdateSolrClient.close()}} gets called twice ... it seems like
> we're just getting really lucky that (as configured by DUP) all of these
> resources are still usable after being finished/shutdown/closed
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
