-
Hi, This is probably a very basic question, but I can't find the answer online. I have a job that failed when uploading outputs. Here is what I see for the logging info:
When I check the Storage Element where the outputs are sent, I see the job completed successfully. However, I cannot just run dirac-wms-job-get-output to recover the log file and find out what exactly went wrong:
It looks as if the temporary folder hosting my job was erased after the "main" job completed but before DIRAC could fetch the outputs. Could I get some help figuring out what is going on here, or an alternative way to access the files that were in the outputsandbox? |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
Replying to my own question here: it looks like the issue was that one of the files in the output sandbox was larger than 100MB. Removing that file from the sandbox (and sending it to storage instead) fixed my problem. Is there any way to get better error messages? I feel like it would have taken much less time for me to figure out the problem if DIRAC had told me why the "uploading outputs" stage failed. |
Beta Was this translation helpful? Give feedback.
Replying to my own question here: it looks like the issue was that one of the files in the output sandbox was larger than 100MB. Removing that file from the sandbox (and sending it to storage instead) fixed my problem.
Is there any way to get better error messages? I feel like it would have taken much less time for me to figure out the problem if DIRAC had told me why the "uploading outputs" stage failed.