action #135509
closedmonitor.qa.suse.de yields 502 Bad Gateway from nginx/1.21.5 size:M
0%
Description
Observation¶
monitor.qa.suse.de yields 502 Bad Gateway from nginx/1.21.5. We looked into logs of the service granafa-server already, couldn't start. Seems to have started just while we looked at it around 2023-09-11 11:00Z
Also https://gitlab.suse.de/openqa/salt-states-openqa/-/jobs/1820468:
ID: grafana-server
1037 Function: service.running
1038 Result: False
1039 Comment: Job for grafana-server.service failed because the control process exited with error code.
1040 See "systemctl status grafana-server.service" and "journalctl -xeu grafana-server.service" for details.
1041 Started: 13:15:54.262049
1042 Duration: 16566.052 ms
1043 Changes:
1044Summary for openqa-monitor.qa.suse.de
Steps to reproduce¶
- Go to https://monitor.qa.suse.de and observe the error
Suggestions¶
- On monitor.qa.suse.de look into
journalctl -u grafana-server
- Just try a plain reboot of the whole VM
- Fix the problem
- Consider improvements to fix in the future
Updated by nicksinger about 1 year ago
- Status changed from In Progress to Feedback
The grafana-server
-service was not able to access /var/run/grafana where grafanas unix socket is created. This was most likely caused because we manage this folder in salt. After a cleanup the folder was not present and salt was not running to create it again before grafana-server
tried to start. I changed our states to create a according systemd-tmpfiles config entry: https://gitlab.suse.de/openqa/salt-states-openqa/-/merge_requests/982
I think systemd-tmpfiles is also responsible for the previous cleanup.
Updated by nicksinger about 1 year ago
- Status changed from Feedback to Resolved
change was deployed and I verified manually that systemd-tmpfiles --create
is only executed on file changes