Closed harper357 closed 7 months ago
Hmm now, looking at the line, this check might actually be performed by nextflow itself and therefore on the head node. But I kind of thought that reading the first X lines should be doable.
This has been open for more than a year now.
We don't index by default anymore and therefore do not check the files anymore. If you provide raw files or indexedMzml, this should not happen anymore.
Description of the bug
Im getting an odd error when trying to run 40 samples through the pipeline on AWS batch.
Everything proceeds normally until the MZMLINDEXING step when the head node crashes, with the error java error “Failed to acquire stream chunk”.
The log where the error happens:
The code refenced:
https://github.com/nf-core/quantms/blob/a1bf7d4104ec424abff984512764ddecde79d21f/subworkflows/local/file_preparation.nf#L30-L34
The head node seems to crash at a different file number if I change the amount of memory I assign to the head node. Is all the data passing through the head node somewhere? I've never had this problem with any of my NGS pipelines, they use more and larger files, so I am a little confused at this crash.
Command used and terminal output
No response
Relevant files
The log file where it crashes: (it is dated different, but this is the same error that always shows)
System information
Nextflow version (eg. 22.04.5) Hardware AWS Executor awsbatch Container engine: default OS AWSLinux Version of nf-core/quantms v1.1dev