I have run my standard CuffDiff workflow which usually finishes overnight. But this time the job is still running for over 24 hours and hasn't finished yet. Should I delete and start again, or just keep waiting?
I have run my standard CuffDiff workflow which usually finishes overnight. But this time the job is still running for over 24 hours and hasn't finished yet. Should I delete and start again, or just keep waiting?
Hello,
If you are using the public Main Galaxy instance at http://usegalaxy.org, then this is how jobs are handled:
http://wiki.galaxyproject.org/Support
Section 2.6 = Dataset status and how jobs execute
Section 2.8.1-4 = 4. Job failure reason "ran longer than the maximum allowed job run time"
Then quotas are described in more detail here, see the table for "Compute resources". Currently all jobs on Main are using the default cluster. This cluster has a quota of 36 hours of wall-time (execution time).
http://wiki.galaxyproject.org/Main
In short, if the job is grey, that would be unexpected at this time unless you have several queued. Jobs will execute in the order submitted - wait for the other jobs to complete and then this job will queue when it is your turn to run a job again (jobs slots are allocated to users evenly). If the job is yellow, the job still has more time to complete running before hitting the "wall-time" for execution. If the job is red because it eventually fails due to exceeding wall-time (or memory), you might be able to adjust parameters or will need to move to a local or cloud Galaxy where you can dedicate more resources.
Hopefully this helps, Jen, Galaxy team