[stable/nextcloud] recommended upgrade method unknown; leads to non-functional nextcloud #17093
Comments
/assign |
So would you mind sharing your nextcloud values? One suggestions from my side would be to use existing secret function of mariadb chart. Then it won´t be created or touched during upgrade. |
is there a way to get the values from kubernetes/helm itself? I'd hate to give you a yml from my laptop which might not represent the full truth. |
If there is no way to get the values from k8s/helm/etcd, then this is the file I initially used:
|
Is the root password getting regenerated and reinstalled somewhere? It doesn't appear to be limited to just upgrading nextcloud. I simply rebooted my machines and I've now got the mariadb containers failing with the same output. My k8s secrets haven't changed since the initial installation a few months ago. Here are my values:
|
So from my perspective it is a point with missing configuration in mariadb values. In mariadb values.yaml there is a block which describes your issue.
So I will close this issue and add forcePassword to the the nextcloud default values, just to be sure. /close |
Nice find. What would be the recommended approach for recovering an instance that has had this happen? Does anyone know of a way to reset the root and replication passwords in the pod? I've tried googling around looking for ways to do this but so far everything I've found is assuming it's running outside of a container and you can kill mysqld. |
Can confirm, btw, if I set forcePassword and rootUser.password and replication.password during the initial install in my values.yaml, I can successfully upgrade an instance from 16.0.3 to 16.0.5 by changing the image tag and running helm upgrade --install nextcloud -f values.yaml, and I can restart the database pods and they will start back up successfully. I tested this on a separate installation. Half tempted to back up the old install and restore to a new deployment, rather than keep trying to fight the container to reset root password. Related: #5167 |
aren't there 2 issues at hand then? I could really bring back my mariadb cluster by looking in the helm history for the values of the secrets, re apply those, and it would resolve itself. |
You can look at the values of past generated secrets? If that's possible, that would fix my install too... edit: yes indeed.
where N is one of the revisions listed in Decode the password:
Paste the password, press enter, and Ctrl+D. Test the passwords in the nextcloud mariadb container:
Once the right one is found:
Paste the base64 encoded data in place, delete the pods. |
This is my values.yaml profile
The deployment cannot be started after the installation is complete, the log shows
The probe tells me
I want to know why I can't connect to port 80? |
@happinesslijian is your issue related to the problem described in the OP ? |
Do you know this problem? Please guide me Value.zip is my configuration file, please use helm to install, you will also see this problem. |
Thanks @ftab for explanation. I think this will help other to recover their installation after upgrade. Additionally more details on that should be addressed to the mariadb helm chart collaborators cause I´m not yet that familiar with the chart. @happinesslijian I don´t think your problem is related to this issue. Any way it is extremely difficult to find out what your problem is if you paste a complete values.yaml. in your values.yaml you only need values that differ from the values.yaml defaults in the chart. With this it would be a lot easier to figure out what might be a problem. I do not really want to go through each property and check if it differs from default values.yaml. Pod logs would also help. And maybe it´s more a topic for a help forum like https://help.nextcloud.com . Thx. |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Any further update will cause the issue/pull request to no longer be considered stale. Thank you for your contributions. |
This issue is being automatically closed due to inactivity. |
Describe the bug
Upgrading to the latest chart version by doing:
Where
nextcloud.yaml
is my values yaml, results in a non functioning nextcloudVersion of Helm and Kubernetes:
Which chart:
stable/nextcloud
What happened:
mariadb comes up but has many failed logins:
consequently, the main pod complains with the following errors:
What you expected to happen:
Nextcloud is upgraded without a hitch
How to reproduce it (as minimally and precisely as possible):
Anything else we need to know:
Seems this is related to secrets being regenerated every time you upgrade. My solution was to rollback, and manually copy the secrets of the mariadb root and replication password. Than re-apply the upgrade, and then edit the secrets by hand to the original value.
The text was updated successfully, but these errors were encountered: