I've been running cockpit for a long while now successfully on an Azure App Service with SQLite, on Windows + IIS + FastCGI.
Recently I have been working on moving the project to a Kubernetes Cluster.
What I did:
Copy Storage and Config to an Azure Storage File Service
Create my StatefulSet based on agentejo/cockpit:0.9.0 (same as the live version)
Mount the File Service via PVC to the cockpit Pod
Give www-data full permissions on /storage and /config
What happens:
First hit goes ok, login dialog opens, I can login.
I can open settings, view collections.
When I try to open a singleton or an entry from a collection, my server request times out. I have 'debug'=>true turned on, but there is no error recorded.
TL;DR:
Basically, opening any "edit" type view which uses the database will crash the app. The config editor works fine. Listing entries in a collection works fine, but not editing an entry.
If I remove the volume mounts and start fresh with /install, everything works fine.
I have no idea where to start troubleshooting this given there is no actual error.
LE: I've ruled out ingress controller issues, exposed the pods via a LoadBalancer Service directly facing the same issue. I've also tried recovering each DB file with sqlite3 ".recover" - no dice.
Weird title, but let me explain:
I've been running cockpit for a long while now successfully on an Azure App Service with SQLite, on Windows + IIS + FastCGI. Recently I have been working on moving the project to a Kubernetes Cluster.
What I did:
agentejo/cockpit:0.9.0
(same as the live version)www-data
full permissions on /storage and /configWhat happens: First hit goes ok, login dialog opens, I can login. I can open settings, view collections. When I try to open a singleton or an entry from a collection, my server request times out. I have
'debug'=>true
turned on, but there is no error recorded.TL;DR: Basically, opening any "edit" type view which uses the database will crash the app. The config editor works fine. Listing entries in a collection works fine, but not editing an entry. If I remove the volume mounts and start fresh with /install, everything works fine.
I have no idea where to start troubleshooting this given there is no actual error.
LE: I've ruled out ingress controller issues, exposed the pods via a LoadBalancer Service directly facing the same issue. I've also tried recovering each DB file with sqlite3 ".recover" - no dice.
My Persistent Volume:
My Stateful Set: