[ 
https://issues.apache.org/jira/browse/SOLR-7180?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14348498#comment-14348498
 ] 

Alan Woodward commented on SOLR-7180:
-------------------------------------

Calling awaitTermination shouldn't be necessary, because invokeAll() won't 
return until everything has finished, but it won't hurt anything and is a good 
check to have.  I'll add it in.

I didn't use stopJettySolrRunner() because it relies on the position of a Jetty 
in the jettys list, which will not be constant if we're calling it in parallel. 
 It's probably worth clearing the list after everything's done though.

Error handling is a slightly tricky one.  Should we fail startup if one of the 
jetties doesn't start?  The behaviour at the moment isn't great - you'll get an 
exception if one of the starting jetties dies, but the already-started-up ones 
will keep running.  At close, I guess the nicest behaviour would be to collect 
all exceptions from the futures, and throw an IOException with any collected 
exceptions added as suppressed.

> MiniSolrCloudCluster should startup and shutdown its jetties in parallel
> ------------------------------------------------------------------------
>
>                 Key: SOLR-7180
>                 URL: https://issues.apache.org/jira/browse/SOLR-7180
>             Project: Solr
>          Issue Type: Improvement
>            Reporter: Alan Woodward
>            Assignee: Alan Woodward
>            Priority: Minor
>         Attachments: SOLR-7180.patch
>
>
> Followup to SOLR-7179.  Now JettySolrRunner doesn't use sysprops to pass 
> configuration, we can start up multiple runners in parallel.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to