Job "config.save" does not finish and constantly yields error 32
Description
Attachments
- 08 Jul 2024, 07:55 PM
Activity
Zack August 16, 2024 at 5:58 PM
@Tor Yes, you could pull it directly in the shell. There might be a way to do it from the UI in ElectricEel with docker container support as shell access has been restricted.
Remove the existing container and recreate it on the new image with the same tied volume and you’re good to go. The command would now be docker pull ghcr.io/ixsystems/truecommand:v3.0.2
after release the other day. After pulling the image, stopping the existing container, removing it (optional), then creating a new one will take a few seconds. Servers should be reconnected in ~30s.
The API page is on the top right, Gear->Admin->API at the bottom. Here you can send API calls through the UI, and the log level can be set accordingly (sys/log-level/set
, {"level":"debug"}
). Another way is to find /etc/middleware/config.yaml
on the container and create a copy outside of the container and deploy a container with a docker volume linked to a local copy. "trace"
is available as well, but that can fill up quickly.
Let me know if you are still able to reproduce the issue.
Tor August 14, 2024 at 9:26 PM
@Zack Where should I run that docker pull command? Directly via System Settings → Shell? And could you elaborate a little bit more on the container section, i.e. when do I notice that the new container has started, which API page etc.? Thanks a million!
Zack July 12, 2024 at 12:41 PM
@Tor Unfortunately I could not reproduce the timeout error and there isn’t a workaround, so the best thing to do now is to update to the latest TrueCommand nightly (docker pull ghcr.io/ixsystems/truecommand:release-3.0.2
) where a fix was attempted (timeout on downloads increased, subsequent downloads will block until the first is complete). Once the new container has started, go to the API page and send {"level":"debug"}
to sys/log-level/set
, then if you reproduce this issue attach the Middleware Log to this ticket and then a call with support can be setup.
Tor July 12, 2024 at 11:20 AM
Is there anything I can do on my end? Like a workaround for the meantime?
The config.save job runs nearly every couple of minutes and it peaks CPU usage but never finishes, cf. screenshot.
Seen in SCALE-24.04.1.1 (DragonFish)