hepcat72
commented
1 year ago Well, I cannot figure out what conditions are causing the hang. I've run it 4 times now and the only time it didn't hang was the attempt right after deleting the output file. 1 subsequent attempt without a pre-existing output file also hung, so now I'm back in debug mode. I did not have this problem for months, so I'm not sure what the issue is. Might not even be idr for all I know at this point. I am going to close this issue. Once I've figure out what the real problem is, I will re-open (if I can isolate it to an issue with idr.
I'm not an experienced user of
idr, but I'm helping to convert an analysis pipeline tosnakemakeand when I was testing on our cluster, I encountered an issue where theidrstep appears to hang. It followed a previous failed snakemake run where the snakemake process had been interrupted. (I'd thought snakemake would have prompted me to remove what had been created as an "incomplete file", but for whatever reason, it didn't. Regardless of what happened, the output file existed beforeidrstarted.) It clearly tried to runidrbecause the STDOUT file contained a single line:So I tried running the same
idrcommand that snakemake had run, manually, and it did the same thing. It output that 1 line to STDOUT and just hung.On a hunch, I checked the existence of the output file, and there was a pre-existing output file. Unfortunately, I didn't look in the file before deleting it. But I deleted it and tried simply running the same idr command again, and this time, it proceeded without any hang.
Is this intended behavior? I would have guessed that idr would either issue an error or overwrite the file. I would not have guessed that it would hang. I tried searching the usage, the issues, and the docs for an explanation of this behavior, but could not find anything relevant.