The http1 and http2 code in the Http2SolrClient ultimately create different
errors when there are TLS issues.
So the fix was merely to accept both types of errors, when checking the
output of failed commands.
These should be fixed on main, branch_9x and branch_9_4.
So let's give it a day and see w
Good idea! To be honest I thought I already had done that awhile back, but
I can't find the email. So maybe it never made it out of my drafts folder
: /
In any case I just started a new thread on users@ about this, titled:
"Feedback/Review Requested: 'Official' Solr Python Client". Hopefully
th
Ok, I've determined that the mTLS errors are due to using http1 vs http2 in
the client.
I'm not sure why http1 is used by default in different environments (Its
enabled by the `-Dsolr.http1=true` System Property),
but now at least the cause is known. I'll try to fix the ssl integration
tests for th
There were a few tests that were failing on PRS seeds. Noble and I fixed
them over the weekend.
There are some intermittent failures on a variety of tests in the branch,
and I'm finding it hard to get a full test run done in order to build the
artifacts. At the moment, none of them look serious or
If you scroll down on the output of
https://builds.apache.org/job/Solr/job/Solr-Check-main/8273, you can see it
in the integration tests.
Search for "not ok 75 start solr with mTLS needed"
For the builds using crave.io (Solr-Check-9.x), the same failures will
happen but every time. The builds usin
