-
Notifications
You must be signed in to change notification settings - Fork 6
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
mysql-router-k8s: One of three units down #173
Comments
I just finished a triage run through of our last set of tests from Solutions QA and I think I can identify 4-5 other test runs that seem to be suffering from the same underlying issue - the mysql-router apps are all active from Juju's perspective but the consuming service has trouble accessing its DB with a connection refused message. |
this makes me think it's related to canonical/mysql-k8s-operator#345 but I'm quite surprised that the juju status is active—did the router recover afterwards? do you have juju debug logs for the from the mysqlrouter.log, it looks like the bootstrap was successful but the cluster went offline after the bootstrap. Not 100% sure though—also if that happened, the mysql-router charm's health check should've shown an error on the next |
it's quite difficult to figure out what happened without the juju debug log |
@javacruft do you have debug logs for these runs? do the routers remain active (i.e. did an update-status fire and they're still active)? is MySQL server active? are you able to connect to the server directly? |
@carlcsaposs-canonical the inspection report has the full debug log for the deployment: |
All of the mysql{-router} units are showing as active: |
thank you! |
okay, I am seeing an error in the debug log
did it recover afterwards? it appears so, but not sure without DEBUG-level logs did it show active status but connections didn't work? |
1.txt |
@javacruft are you able to reproduce this with debug-level logs? while this issue initially looks similar to canonical/mysql-k8s-operator#345, on a closer look, it seems like it might be a different issue what's surprising to me is this part of the logs
it seems like a transient issue that went away on a retry, but it sounded like the router was showing active status with the router configuration file missing—if it's the second case, debug logs would be quite helpful |
Likely fixed by #187. Please re-open if you're able to reproduce or if you have debug-level logs |
Steps to reproduce
Failed multi-node test run from Canonical Solutions QA team.
Multi-node microstack deployment on baremetal with deployment in many-mysql mode - mysql per service.
Then enabled telemetry with
sunbeam enable telemetry
All other mysql and mysql-router apps deploy and scale correctly however one mysql-router-k8s instance failed to bootstrap.
Expected behavior
All mysql-router-k8s units bootstrap correctly.
Actual behavior
Failure of single mysql-router-k8s unit. This is not reflected in the workload status of the unit:
Versions
Operating system: 22.04.3
Juju CLI: 3.2.4-genericlinux-amd64
Juju agent: 3.2.4
mysql-k8s charm revision: 99
mysql-router-k8s charm revision: 69
microk8s: 1.28-strict/stable
Log output
Juju debug log:
Log from failed run: https://oil-jenkins.canonical.com/artifacts/e149fed6-4aee-4a4a-b6ee-d6d931a503f5/index.html
Additional context
This is very similar to issue canonical/mysql-k8s-operator#345 but in this case there is no mysql-router hook error, the mysql router bootstrap appears to have silently failed,
The text was updated successfully, but these errors were encountered: