Hey has an alternate wip branch that is a bit more up to date, but still 9
different fails.
He also has a benchmark for the benchmark module - that will help with
testing, but we will definitely want to bang on that issue before release.
I doubt that banging should hold up branching though - I’m s
Hi
I observed after upgrading solar cluster to 7.7 from 6.5. I observed solr
nodes going down with high cpu spike after using for a day, I also upgraded
java version from Open JDK 8 to Open JDK 11 while upgrading solar cluster.
There seems to be different GC algorithm used by Java 11 I don’t
If anyone is interested in testing the Solr operator against the
recently cut Solr 8.10 branch, I just pushed a Docker image (built
locally) to: thelabdude/apache-solr-dev:8.10.0-SNAPSHOT
SolrCloud CR YAML:
solrImage:
repository: thelabdude/apache-solr-dev
tag: 8.10.0-SNAPSHOT
Cheers,
Thank you for the updates Andrzej (via slack), Mike and Mark ... I'll
wait to hear back from Mike D on 1 before proceeding with the RC
...
On Wed, Sep 8, 2021 at 11:12 AM Mark Miller wrote:
>
> I see there is alternate more recent SOLR-1-wip branch i missed. Taking a
> look at that.
>
>
I see there is alternate more recent SOLR-1-wip branch i missed. Taking
a look at that.
MRM
On Wed, Sep 8, 2021 at 2:34 AM Mark Miller wrote:
> SOLR-1 is critical query path and appears able to currently fail up to
> 150 tests per run due to what looks mostly to be stats/metrics races,
OK LUCENE-10088 looks like only an exotic random test issue (thank you
Adrien for understanding what the test was doing, to hit "too many open
files"), not a blocker for 8.10.
Mike McCandless
http://blog.mikemccandless.com
On Wed, Sep 8, 2021 at 3:34 AM Mark Miller wrote:
> SOLR-1 is crit
Follow up on the Solr docker tests build failure. It seems that the job
succeeds when scheduled on lucene-1 and fails with "setfacl command is
unavailable" when scheduled on lucene-2.
I created a ticket with infra to make sure that the "acl" package is
installed on both build machines. INFRA-22305
SOLR-1 is critical query path and appears able to currently fail up to
150 tests per run due to what looks mostly to be stats/metrics races,
though it’s hard to be sure that’s all with all the noise. Do you have an
update you can push Mike?
MRM
On Tue, Sep 7, 2021 at 12:35 PM David Smiley w