This morning I launch an update on my internal gitlab-ee instance from 15.9.3 to 15.11.8 and I reach an “ERROR 500 something goes wrong” on the web interface at the end.
My OS is CentOS7 up-to-date (ran a yum update --exclude gitlab-ee before) and I juste ran:
yum install gitlab-ee-15.11.8-ee.0.el7
In the logs I had an error:
RuntimeError: redis_service[redis] (redis::enable line 19) had an error…
Nor a reboot, nor a gitlab-ctl reconfigure provides any help.
Searching for help, I read it could be related to selinux which is enabled on my server. So I switch selinux to Permissive before launching again reconfigure:
setenforce Permissive gitlab-ctl reconfigure
This launch a long database update that should have failed before and seams to solve the problem.
Then I have rebooted the server, Selinux went back to enabled and all seams working.
May be there is something wrong in the rpm update scripts to investigate.
I have a second server running Alma Linux8 to update, may be I will set selinux to Enforcing before running the update…