"Too many open files" error

[wuyilei]

when running nextflow on a grid engine cluster, always got a "too many open files" error during the middle of the running. detailed log as follows.

$ cat .command.log
finding positions of the duplicate reads in the file...
sambamba-markdup: Cannot open or create file './sambamba-pid6283-ztgh/sorted.50.bam' : Too many open files

is it possible related to nextflow?

[pditommaso]

This is not directly related to nextflow. That error is reported by the sambamba-markdup tool. If you are using nextflow to launch many instances of that tool it could happen that you reached the maximum amount of files that can be opened in your system.

The easiest solution is to increase that limit with ulimit. See, for example, here http://askubuntu.com/questions/162229/how-do-i-increase-the-open-files-limit-for-a-non-root-user

[pditommaso]

I'm closing the issue. Feel free to comment if you need more explanations.

[wuyilei]

unfortunately, our ulimit is set to unlimited

[pditommaso]

That is weird. Are you sure you are checking that in the cluster computing
node?

[wuyilei]

i found the following link may be helpful to solve it.

biod/sambamba#177

[wuyilei]

yes, we check ulimit on every node of the cluster

[pditommaso]

Yes, this confirms that it's a problem with sambamba (I mean the biod/sambamba#177 issue)