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
k8s-csi-cleaner empty string error #301
Comments
Can you |
|
Do any interesting logs show up from the respective |
Apologies for letting this linger. Even after rebuilding on TrueNAS-SCALE-Cobia (latest release), ie importing the zpools onto a fresh install, this is still an issue. Current environment: The error (triggered in same way) seems slightly different.
To very belatedly answer your question, that command results in the following
|
Can you turn on debug logging? It will show the the actual request/response from the TrueNAS API. |
This is an example response from CI:
If you search |
Debug logging output (from nfs-controller/csi-driver, using aforementioned versions) is here |
I can't get that link to open :( |
Think that pastebin choked on syntax highlighting the (lengthy) debug response. |
hmm, I don't see that error in the output. Can you recreate the error and send over the respective logs? |
Apologies again, time to lab is fleeting at present :( Debug output captured from [2024-02-12 edit: pastebin url without expiry] |
We need to turn on debug logging so you see the raw requests/responses :( this issue seems really odd. I can add a small snippet of code to maybe try and catch it with more robust detail. |
My fault! Had |
Well that response seems to have completed successfully. Can you run those logs out further to the point where the cleaner breaks? |
Make sure to run the cleaner from the controller pod if that's not clear. |
:( Sorry this is so back and forth, really not sure what I'm doing wrong to upload healthy logs when the the error is so repeatable. Will try to be verbose in hope of it helping you identify my misstep:
Relevant 5min timespan of the entire democratic-csi/democratic-csi-nfs-controller If I can provide more context, logs, anything, please do let me know. |
Switching from My Is it possible |
No. I looked through the previous logs sent and still cannot find a response where the children attribute is not present. Crazy stuff. If you actually did the clean operation with the ssh config have you tried going back to the api config and testing if it still blows up? |
Just tried, no dice.
If there are no significant downsides to using the SSH reliant drivers over the API-only ones (with TrueNAS SCALE), perhaps this working configuration can be considered a resolution? |
There are no downsides if you don't mind allowing ssh access to the node. I would like to discover the issue. The last set of logs you sent do not include a respective |
Can you also send over your full config (cleansed of secrets of course)? |
K8s distro = Talos 1.4.x
TrueNAS distro = SCALE Bluefin (22.12.2-1)
Pod: democratic-csi/csi-driver
Has been true for a while (many minor Talos/SCALE revisions), just been lazy with logging the fault.
This is particularly painful when paired with the lack of an
onDelete
param for the storageClass; many dangling child datasets, no programmatic way to clean them up.Happy to provide more info/logs, please advise!
The text was updated successfully, but these errors were encountered: