You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@solr.apache.org by "Chris M. Hostetter (Jira)" <ji...@apache.org> on 2023/05/25 23:16:00 UTC
[jira] [Updated] (SOLR-16819) (nightly) HttpPartitionTest broken regardless of seed
[ https://issues.apache.org/jira/browse/SOLR-16819?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Chris M. Hostetter updated SOLR-16819:
--------------------------------------
Attachment: SOLR-16819.patch
Status: Open (was: Open)
Most of the changes made in SOLR-10466 look like this...
{noformat}
- cloudClient.setDefaultCollection(testCollectionName);
- sendDoc(1);
+ sendDoc(testCollectionName, 1);
{noformat}
Which made this change stick out like a sore thumb...
{noformat}
- cloudClient.setDefaultCollection(testCollectionName);
-
sendDoc(1);
{noformat}
Attached patch adresses this, removing the one arg {{sendDoc(int)}} to ensure nobody ever makes that mistake in any future test edits.
----
FWIW: the section of the test that fails is gated like this...
{code}
if (TEST_NIGHTLY) {
// each time
testRf3();
}
{code}
...i'm guessing that when you were "...currently working on making sure the nightly tests pass..." you did something like this...
{noformat}
-Ptests.filter='@Nightly'
{noformat}
...to beast the nightlies while ignoring the other tests you had already fixed. But that only ensures that tests with the \@Nightly annotation get run -- it doesn't impact things like {{TEST_NIGHTLY}} or {{TEST_MULTIPLER}} the way {{-Ptests.nightly=true}} does.
So you probably saw failures from other lines in the test, fixed those, and not noticed that these lines also needed fixed because they wouldn't have caused failures w/o {{-Ptests.nightly=true}} being explicitly set.
> (nightly) HttpPartitionTest broken regardless of seed
> ------------------------------------------------------
>
> Key: SOLR-16819
> URL: https://issues.apache.org/jira/browse/SOLR-16819
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Chris M. Hostetter
> Priority: Major
> Attachments: SOLR-16819.patch
>
>
> {{HttpPartitionTest}} (which is a nightly test) has been failing 100% of the time on jenkins for the past month (2023-05-01 to 2023-05-25 at time of writting)
> Before that – we have no reported failures since 2022-07, which is weird because up to that point this test was always a bit flakey.
> Checking out older version of main from april and running {{./gradlew -p solr/core test -Ptests.nightly=true --tests HttpPartitionTest}} show that this test definitely *should* have failed a lot of nightly builds prior to 2023-05-01 ... so i'm assuming we had some jenkins or gradle build glitch that prevented it from being run? (i didn't dig into this)
> picking an arbitrary commit point from a few months ago when i could confirm the test *did* pass i then ran {{git bisect}} and it identified 708524315bf64ed5dbcfd470c8fadd435c861f78 / SOLR-10466 as the first bad commit (which is also the most recent commit to this file)
>
> (NOTE: this problem also seems to affect the subclass {{HttpPartitionWithTlogReplicasTest}} which has also been failing 100% of the time for the past month)
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@solr.apache.org
For additional commands, e-mail: issues-help@solr.apache.org