Fixes for test flakes #12775
Fixes for test flakes #12775
Conversation
The test checks later based on clientId
dcorbacho
force-pushed
the
fix-flakes
branch
from
d5e0b41
to
c840c11
Compare
dcorbacho
force-pushed
the
fix-flakes
branch
from
c840c11
to
fabf2e9
Compare
Parallel/sharding groups often fail to create certificates in CI. Most likely it is related to the fact they use the same directory for certificates. This commit uses a different directory per group
initializes rabbitmq configuration
To capture where the flake occurs
| @@ -281,11 +285,8 @@ init_per_testcase(TestCase, Config0) -> | |||
|
|
|||
| case ?config(group, Config) of | |||
| https -> | |||
| start_https_oauth_server(?AUTH_PORT, ?config(rmq_certsdir, Config), | |||
| start_https_oauth_server(?AUTH_PORT, ?config(certsdir, Config), | |||
There was a problem hiding this comment.
I don’t understand what is the initial problem here. Now groups copy rmq_certsdir to certsdir: why didn’t it work with rmq_certsdir in the first place?
| Group = proplists:get_value(name, | ||
| ?config(tc_group_properties, Config)), |
There was a problem hiding this comment.
This function is often called from init_per_suite/1 where the group is not set. proplists:get_value/2 will return undefined, so this really won’t solve the problem. What about using the VM PID perhaps? Assuming the sharding is based on running multiple VMs, but I don’t know how it works :-)
There was a problem hiding this comment.
Good catch. The most common flake has groups so I hadn't realised. I'll look into it
|
I cherry-picked 7e1cb05 to main since it seems to reliably fix the "Authentication/Authorization backends via mutiple messaging protocols / selenium" job. |
Just another batch of fixes for test flakes