Open ernstae opened 1 year ago
I wouldn't call it a bug for the simple reason that the way the data container is created does not follow what our docs recommend.
The error app provisioning error: plugin not installed: "pmm-app"
means that PMM could not find the contents of the plugin and therefore failed to provision it.
One thing I highly recommend is to provision the data volume following our instructions and possibly remove the ln -s command
from the pmm-server
container - it should solve it.
another lifehack you could use is to pass GF password/user for admin as an env var GF_SECURITY_ADMIN_USER
: https://github.com/percona/percona-helm-charts/blob/main/charts/pmm/templates/statefulset.yaml#L78
Or as a file (although I didn't verified that): https://grafana.com/docs/grafana/latest/setup-grafana/configure-docker/#configure-grafana-with-docker-secrets
GF_SECURITY_ADMIN_PASSWORD__FILE=/run/secrets/admin_password
Description
I have yet to run the container image for 2.33.0 successfully. I've tested upgrading (using my data container + changing the server image which has worked flawlessly in the past). I've also tested a new installation and at this point, haven't been able to successfully start the pmm-server containers. The issue appears to occur at this point of the instantiation:
pmm-server | 2023-01-26 06:21:04,252 INFO exited: qan-api2 (exit status 1; not expected)
and grafana goes into a crashloop cycle after reporting aExpected Results
Service starts correctly and presents login screen
Actual Results
500 Error in Nginx (bad gateway).
Resulting in unhealthy containers
All pointing to grafana not loading the pmm-app plugin
Version
PMM Server 2.33.0
Steps to reproduce
Using this docker-compose.yaml file:
Relevant logs
Code of Conduct