I also built a test docker image (using the command above) and ran it with
the Solr Operator integration tests (check out the solr operator repo and
run "make e2e-tests SOLR_IMAGE="solr-rc:9.2.0-1").

Got a success!

Ran 6 of 6 Specs in 279.874 seconds
SUCCESS! -- 6 Passed | 0 Failed | 0 Pending | 0 Skipped


Ginkgo ran 1 suite in 4m58.443616072s
Test Suite Passed

I was having trouble with this.
>

Can you make a JIRA/separate issue about this Shawn? I had my own issue
with the integration tests running very slowly that I'm going to split off
into a separate issue.

- Houston

On Wed, Mar 22, 2023 at 1:53 PM Shawn Heisey <apa...@elyograg.org> wrote:

> On 3/21/23 09:25, Houston Putman wrote:
> > You can run the smoke tester directly with this command:
> >
> > python3 -u dev-tools/scripts/smokeTestRelease.py \
> >
> https://dist.apache.org/repos/dist/dev/solr/solr-9.2.0-RC1-rev-92c5515e2918c22513f7aa527d6c6db943150fea
>
> +1 from me as I was eventually able to get a successful smoke test.
>
> ----------------------------------------
>
> I was having trouble with this.
>
> I discovered that it puts a lot of data into /tmp.  On my primary
> server, /tmp is a 4GB ramdisk (type tmpfs in Linux) as I mentioned on my
> wiki page.  Search for "four month" to find that info:
>
> https://cwiki.apache.org/confluence/display/SOLR/ShawnHeisey
>
> When I switched to a different system that has no separate /tmp mount,
> the smoke test passes every time.  On the system where it worked, I
> never saw the size of /tmp get above 2.9GB, but that was checking every
> 10 seconds, so maybe it got above that briefly, or failed to do
> something because it could see there wasn't enough space.
>
> I do not know whether the problem is available space in my /tmp ramdisk
> or maybe a problem doing hardlinks.  But it strikes me as a bad thing
> that the smoketester script uses /tmp for large data.  I tried another
> test on that second system where the git checkout was on a different
> filesystem than the root (where /tmp lives) and it still passed, so a
> hardlink problem does not seem likely.
>
> ---
>
> I think the temporary directory for unpacking the release candidate and
> running tests should be created as a subdirectory of the source tree,
> not in /tmp.  And then everything except for the test log should be
> cleaned up afterwards whether the smoke test passes or fails.
>
> Thanks,
> Shawn
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscr...@solr.apache.org
> For additional commands, e-mail: dev-h...@solr.apache.org
>
>

Reply via email to