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
At the freebayes step there's a message from parallel that there are no more file handles:
parallel: Warning: No more file handles.
parallel: Warning: Raising ulimit -n or /etc/security/limits.conf may help.
While that's usually not a problem (i.e., it will still continue), it sometimes needs to wait for a specific job to finish. If there's a single contig that takes a long time (some took several hours with one of the datasets), it will wait for this and thus slow down the whole process significantly (usually a single contig takes only a few seconds). Not sure if increasing the file handles is per se a good option, but maybe if would help to set -g/--skip-coverage to something like 100'000 or 1'000'000 (maximal coverage - more than 1000 per sample likely indicates a repetitive element and more than 1000 samples are unlikely as well - you could also set it to numSamples*1000). Also --min-coverage could probably be increased to numSamples or alike as it would speed up the process and just remove contigs that are usually filtered afterwards anyway.
Hi Maarten
At the freebayes step there's a message from parallel that there are no more file handles:
parallel: Warning: No more file handles.
parallel: Warning: Raising ulimit -n or /etc/security/limits.conf may help.
While that's usually not a problem (i.e., it will still continue), it sometimes needs to wait for a specific job to finish. If there's a single contig that takes a long time (some took several hours with one of the datasets), it will wait for this and thus slow down the whole process significantly (usually a single contig takes only a few seconds). Not sure if increasing the file handles is per se a good option, but maybe if would help to set -g/--skip-coverage to something like 100'000 or 1'000'000 (maximal coverage - more than 1000 per sample likely indicates a repetitive element and more than 1000 samples are unlikely as well - you could also set it to numSamples*1000). Also --min-coverage could probably be increased to numSamples or alike as it would speed up the process and just remove contigs that are usually filtered afterwards anyway.
Or maybe this solution here:
https://unix.stackexchange.com/questions/624952/parallel-warning-no-more-file-handles
Best,
Marc
The text was updated successfully, but these errors were encountered: