-resume does not work since SplitIntervals

[qiyubio]

Hi, we are having similar issues as #30, and I already set up cache='lenient' as show bellow, but when I resume the pipeline, it has different cache hash:

grep SplitIntervals .nextflow.log

Jan-27 13:36:17.448 [main] DEBUG nextflow.script.ProcessConfig - Config settings `withLabel:nextNEOpiENV` matches labels `nextNEOpiENV` for process with name SplitIntervals
Jan-27 13:36:17.650 [Actor Thread 16] INFO  nextflow.processor.TaskProcessor - [SplitIntervals (SplitIntervals)] cache hash: 5c88a55030fbe6c561f0f6523481fe7a; mode: LENIENT; entries:

grep SplitIntervals .nextflow.log.1

Jan-27 13:24:31.319 [main] DEBUG nextflow.script.ProcessConfig - Config settings `withLabel:nextNEOpiENV` matches labels `nextNEOpiENV` for process with name SplitIntervals
Jan-27 13:24:31.506 [Actor Thread 13] INFO  nextflow.processor.TaskProcessor - [SplitIntervals (SplitIntervals)] cache hash: 3518123285c713373234e88edeaed7fe; mode: LENIENT; entries:

I've check the checksum is the same from two 0000-scattered.interval_list in each work directory, and the same as the fullOutput. So I don't know why it re-calculates again.

md5sum /vf/users/y3/data_test/nextNEOpi/sm/work/f0/df4afbb6c29bc15b06bbf1a3db756f/wgs_ScatterIntervalsByNs/0000-scattered.interval_list
467830eb79a035cafdca339fd98f359b /vf/users/y3/data_test/nextNEOpi/sm/work/f0/df4afbb6c29bc15b06bbf1a3db756f/wgs_ScatterIntervalsByNs/0000-scattered.interval_list

md5sum /vf/users/y3/data_test/nextNEOpi/sm/work/86/b8f19903aa28377ff7d4008da198b5/wgs_ScatterIntervalsByNs/0000-scattered.interval_list
467830eb79a035cafdca339fd98f359b /vf/users/y3/data_test/nextNEOpi/sm/work/86/b8f19903aa28377ff7d4008da198b5/wgs_ScatterIntervalsByNs/0000-scattered.interval_list

md5sum test_results_sm/supplemental/00_prepare_Intervals/SplitIntervals/wgs_ScatterIntervalsByNs/0000-scattered.interval_list
467830eb79a035cafdca339fd98f359b test_results_sm/supplemental/00_prepare_Intervals/SplitIntervals/wgs_ScatterIntervalsByNs/0000-scattered.interval_list

Any thoughts?

[riederd]

Hi, I'm trying to reproduce the issue, is this with WES or WGS data, what filesystem do you use?

[qiyubio]

Only happens when "WES= false", seems related to

nextNEOpi/nextNEOpi.nf

Line 666 in 52d7b92

outFileName = (params.WES) ? interval_list.baseName + "_merged_padded.interval_list" : "wgs_ScatterIntervalsByNs.interval_list"

. The program was running on HPC/NFS.