[
https://issues.apache.org/jira/browse/SOLR-15172?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17412405#comment-17412405
]
Uwe Schindler edited comment on SOLR-15172 at 9/9/21, 7:34 AM:
---------------------------------------------------------------
Hi, I answered in Slack:
{quote}
I would not separate the Lucene nodes for several reasons: The 2 VMs are
"standard" jenkins VMs, we just have them separate, because Lucene and Solr
jobs are running all the time in a loop (for better randomized testing). As we
trigger jobs to fill those 2 machines 100% we have them under a separate label.
If we would use the general Jenkins pool, we would fill nodes used by other
projects and possibly also many Jenkins jobs would run in parallel on many
nodes. As Solr and Lucene use more or less the same "test framework", the usage
pattern is similar. Also builds of 8.x are combined builds anyway, Solr and
Lucene are one job. If we would split nodes the following would get unweighted:
Where to execute the 8.x jobs? If we execute on one node, the other node is
underutilized! So my suggestion is to keep it as is. The Nodes are maintained
by ASF, and for setup reasons I have root access on them. Nothing more. It's
just a naming problem.
So we can rename the node to LuceneSolr1 and LuceneSolr2 and maybe also put
them in a renamed label "lucenesolr" instead of "lucene". But this may require
to touch all jobs to update the label.
{quote}
was (Author: thetaphi):
Hi, I answered in Slack:
{quote}
I would not separate the Lucene nodes for several reasons: The 2 VMs are
"standard" jenkins VMs, we just have them separate, because Lucene and Solr
jobs are running all the time in a loop (for better randomized testing). As we
trigger jobs to fill those 2 machines 100% we have them under a separate label.
If we would use the general Jenkins pool, we would fill nodes used by other
projects and possibly also many Jenkins jobs would run in parallel on many
nodes. As Solr and Lucene use more or less the same "test framework", the usage
pattern is similar. Also builds of 8.x are combined builds anyway, Solr and
Lucene are one job. If we would split nodes the following would get unweighted:
Where to execute the 8.x jobs? If we execute on one node, the other node is
underutilized! So my suggestion is to keep it as is. The Nodes are maintained
by ASF, and for setup reasons I have root access on them. Nothing more. It's
just a naming problem.
{quote}
> Setup Jenkins for new solr repo build
> -------------------------------------
>
> Key: SOLR-15172
> URL: https://issues.apache.org/jira/browse/SOLR-15172
> Project: Solr
> Issue Type: Sub-task
> Reporter: Jan Høydahl
> Assignee: Uwe Schindler
> Priority: Major
> Fix For: main (9.0)
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@solr.apache.org
For additional commands, e-mail: issues-h...@solr.apache.org
