Closed philsteinberg closed 1 year ago
My best suggestion right now is to just resubmit and see if the error goes away. The cluster has been slightly unstable lately. It seems like your work_dir
change in the config was successful because there is stuff written there.
Maybe @yashpatel6 has seen this error elsewhere?
Update: Running. Re-submitted a bunch and they seem to be running and not erroring with exit code 1 immediately.
That error is usually caused by an issue with the /scratch
disk not connecting properly to the node; there's not much that can be done other than re-submitting
@philsteinberg do you happen to have the job ID for the one that failed?
@yashpatel6 pretty much most between 13831-13795 on Sat between 10:35am-12:45pm and then previously 13707-13576 Fri between 1am and 5am
Several re-runs are completing.
New Errors for samples:
Error log: /hot/project/method/AlgorithmEvaluation/BNCH-000082-SRCRNDSeed/pipeline-call-src/run-mutect2-battenberg-phylowgs/logs/ILHNLNEV000004-T001-P01-F_13142_Mutect2-Battenberg-PhyloWGS.log
Error message:
Error executing process > 'workflow_phylowgs:write_results_PhyloWGS (1)'
Caused by:
Process `workflow_phylowgs:write_results_PhyloWGS (1)` terminated with an error exit status (1)
Command executed:
set -euo pipefail
python2 /phylowgs/write_results.py --include-ssm-names ILHNLNEV000004-T001-P01-F trees.zip PhyloWGS-2205be1_13142_ILHNLNEV000004-T001-P01-F_Mutect2-Battenberg-summ.json.gz PhyloWGS-2205be1_13142_ILHNLNEV000004-T001-P01-F_Mutect2-Battenberg-muts.json.gz PhyloWGS-2205be1_13142_ILHNLNEV000004-T001-P01-F_Mutect2-Battenberg-mutass.json.gz
Command exit status:
1
Command output:
(empty)
Command error:
Traceback (most recent call last):
File "/phylowgs/write_results.py", line 58, in <module>
main()
File "/phylowgs/write_results.py", line 50, in main
munger.remove_multiprimary_trees(args.max_multiprimary)
File "/phylowgs/pwgsresults/result_munger.py", line 96, in remove_multiprimary_trees
len(self._tree_summaries)
Exception: 100% of trees are multiprimary (2500 of 2500), so not enough to report good posterior.
Work dir:
/hot/project/method/AlgorithmEvaluation/BNCH-000082-SRCRNDSeed/pipeline-call-src/run-mutect2-battenberg-phylowgs/debug/c6/5c2e3d30916d668ed8f3fcee3fb4f3
Tip: when you have fixed the problem you can continue the execution adding the option `-resume` to the run command line
executor > local (6)
[a6/90edff] process > run_validate_PipeVal (2) [100%] 3 of 3 ✔
[0b/0f0d63] process > create_inputs_SRCutil (1) [100%] 1 of 1 ✔
[- ] process > workflow_pyclonevi:fit_mode... -
[- ] process > workflow_pyclonevi:write_re... -
[72/f8a051] process > workflow_phylowgs:call_mult... [100%] 1 of 1 ✔
[c6/5c2e3d] process > workflow_phylowgs:write_res... [100%] 1 of 1, failed: 1 ✘
[- ] process > workflow_phylowgs:index_dat... -
[- ] process > workflow_dpclust:generate_i... -
[- ] process > workflow_dpclust:call_RunDP... -
[- ] process > workflow_pyclone:run_analys... -
[- ] process > workflow_fastclone:run_solv... -
Error executing process > 'workflow_phylowgs:write_results_PhyloWGS (1)'
Caused by:
Process `workflow_phylowgs:write_results_PhyloWGS (1)` terminated with an error exit status (1)
Command executed:
set -euo pipefail
python2 /phylowgs/write_results.py --include-ssm-names ILHNLNEV000004-T001-P01-F trees.zip PhyloWGS-2205be1_13142_ILHNLNEV000004-T001-P01-F_Mutect2-Battenberg-summ.json.gz PhyloWGS-2205be1_13142_ILHNLNEV000004-T001-P01-F_Mutect2-Battenberg-muts.json.gz PhyloWGS-2205be1_13142_ILHNLNEV000004-T001-P01-F_Mutect2-Battenberg-mutass.json.gz
Command exit status:
1
Command output:
(empty)
Command error:
Traceback (most recent call last):
File "/phylowgs/write_results.py", line 58, in <module>
main()
File "/phylowgs/write_results.py", line 50, in main
munger.remove_multiprimary_trees(args.max_multiprimary)
File "/phylowgs/pwgsresults/result_munger.py", line 96, in remove_multiprimary_trees
len(self._tree_summaries)
Exception: 100% of trees are multiprimary (2500 of 2500), so not enough to report good posterior.
Work dir:
/hot/project/method/AlgorithmEvaluation/BNCH-000082-SRCRNDSeed/pipeline-call-src/run-mutect2-battenberg-phylowgs/debug/c6/5c2e3d30916d668ed8f3fcee3fb4f3
Tip: when you have fixed the problem you can continue the execution adding the option `-resume` to the run command line
Since the quality of the posterior distribution depends on the number of MCMC chains, I am guessing that increasing the --num-chains 1
would probably fix this issue. However, this does not make sense in our case because that requires inputting more seeds. Would increasing --mcmc-samples 2500
help?
I also found a a post with a similar issue where the suggestion was to add--allow-polyclonal --include-polyclonal
, but I cannot find these flags in the PhyloWGS documentation.
Exception: 100% of trees are multiprimary (2500 of 2500), so not enough to report good posterior.
You are right the crux of the issue is here, basically PhyloWGS is finding solutions of multiple primaries (independent tumours arising in the same organ that are not from the same ancestor clone). When we initially tested PhyloWGS, we found this solution highly unlikely at the rate that PhyloWGS was calling it, and they decided to ban it. Essentially this just means that PhyloWGS failed for this sample + seed.
We can just make a note that it failed, or if you want to be thorough, you can try digging for the appropriate flag in their code and adding it, just so that we save the multiple primary solution (they call it polyclonal, when we say polyconal we mean something else).
Slurm Job_id=17380 Name=ILHNLNEV000011-T001-P01-F_659767_Mutect2-Battenberg-PhyloWGS Failed, Run time 19:59:21, NODE_FAIL, ExitCode 0
- Wed 21/12/2022 @yashpatel6 in case you wanted the data on recent node-fails.
Slurm Job_id=17380 Name=ILHNLNEV000011-T001-P01-F_659767_Mutect2-Battenberg-PhyloWGS Failed, Run time 19:59:21, NODE_FAIL, ExitCode 0
- Wed 21/12/2022 @yashpatel6 in case you wanted the data on recent node-fails.
Got it, thanks!
@lydiayliu Some of the runs have completed and currently are running successfully. However, over night there have been a bunch that failed immediately.
Example failed sample:
ILHNLNEV000014-T001-P01-F_628019_Mutect2-Battenberg-PhyloWGS Failed
Example .error (.log is empty):
/hot/project/method/AlgorithmEvaluation/BNCH-000082-SRCRNDSeed/pipeline-call-src/run-mutect2-battenberg-phylowgs/logs/ILHNLNEV000014-T001-P01-F_628019_Mutect2-Battenberg-PhyloWGS.error
Error message:
mktemp: failed to create directory via template ‘/scratch/XXXXXXX’: Permission denied
This is strange because I changed the
work_dir
in the config from scratch to:/hot/project/method/AlgorithmEvaluation/BNCH-000082-SRCRNDSeed/pipeline-call-src/run-mutect2-battenberg-phylowgs/debug
It also confuses me why this is an error for some but not other runs.