You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I run concourse via the quick-start method under docker-compose, on a single machine. The machine ran out of disk space on its docker volume, and jobs started failing with no space left on device. I freed up space on the device and restarted concourse (via docker restart <concourse container id>).
Now, whenever I restart any job, I get:
initializing
file not found
errored
The job previously ran without this error, and the job has not been changed since it worked. I noticed that the job begins with a git step, which works without issue, and the next step is a run, which is where file not found occurs.
Steps to Reproduce
existing pipeline using git resource followed by run
fill space up on Docker host, until new jobs start failing with no space left on device
free up space and restart concourse
new jobs fail with file not found
Expected Results
Any corrupt state caused by the disk filling up should be corrected by concourse transparently so that workers/jobs can execute normally after recovering from this error.
Actual Results
New jobs fail with file not found.
Additional Context
Concourse 5.0.0
Version Info
Concourse version: 5.0.0
Deployment type Docker
Infrastructure/IaaS: Arch Linux rolling
Did this used to work? I hit a similar thing under 4.x; clearing the database and re-initialising concourse (adding users and pipelines from scratch) was the only way to clear out corrupt state and get things working again.
The text was updated successfully, but these errors were encountered:
Hmm, hopefully not. I ran docker volume prune while concourse was still running. Note that the "file not found" errors were occurring before I did this; the error message didn't change before/after I ran the prune.
Bug Report
I run concourse via the quick-start method under
docker-compose
, on a single machine. The machine ran out of disk space on its docker volume, and jobs started failing withno space left on device
. I freed up space on the device and restarted concourse (viadocker restart <concourse container id>
).Now, whenever I restart any job, I get:
The job previously ran without this error, and the job has not been changed since it worked. I noticed that the job begins with a
git
step, which works without issue, and the next step is arun
, which is wherefile not found
occurs.Steps to Reproduce
git
resource followed byrun
no space left on device
file not found
Expected Results
Any corrupt state caused by the disk filling up should be corrected by concourse transparently so that workers/jobs can execute normally after recovering from this error.
Actual Results
New jobs fail with
file not found
.Additional Context
Version Info
The text was updated successfully, but these errors were encountered: