[BUG] "WARNING Final process status is permanentFail" with pgap-2022-04-14.build6021

[wanjinhu]

hello there:

When i run pgap.py with "test_genome/MG37/input.yaml", but finished with "WARNING Final process status is permanentFail" and there are no valued result.

Apparently I set the --cpus parameter, based on https://github.com/ncbi/pgap/issues/59. But still there is a “WARNING Final process status is permanentFail” problem.

I am using a linux system of centos7. I installed pgap based on the requirements of https://github.com/ncbi/pgap/wiki/Installation. Because of network download and other reasons, I manually installed the corresponding version of Docker image, Supplemental reference data, and CWL code. The versions of these three parts are all "2022-04-14.build6021". And i modified the part of pgap.py, mainly about the update part.

I provided the cwltool.log log file. Based on the cwltool.log, I don't think the problem WARNING Final process status is permanentFail is the reason for modifying pgap.py.

Here are log file: cwltool.log

I'd appreciate it if anyone could help

thanks

[azat-badretdin]

Thank you for your report. The complaint is about the format of the input file. Most likely you are using an old example from the documentation:

ncbi::CObjectIStreamJson::UnexpectedMember() --- line 1: "taxon": unexpected member, should be one of: "strain" "genus_species"  ( at JsonValue.organism)

[wanjinhu]

Thank you for your reply.

I have rearranged the input file as https://github.com/ncbi/pgap/wiki/Input-Files required, but the "WARNING Final process status is permanentFail" problem still occurs, and I still can't solve the problem.

Metadata YAML file (submol) this input I removed all optional content.

Here are other files i used:

The log file: cwltool.log

I think this might be the reason for the failure, but not sure:

pgap/pgap.cwl:22:7: Warning: Field `location` contains undefined reference to
                    `file:///pgap/pgap/input`

thanks again

[wanjinhu]

Hi there,

I still can not fix the problem of this issue (the second comment). I checked the cwltool.log again and found the input information summarize:

    "report_usage": true,
    "no_internet": true,
    "gc_assm_name": "my_gc_assm_name",
    "make_uuid": true,
    "blast_hits_cache_data": null,
    "ignore_all_errors": null,
    "uuid_in": null

And i checked the pgap.cwl, input required in line 23-39 like this:

  blast_hits_cache_data:
    type: Directory?
  fasta: File
  gc_assm_name: 
    type: string
    default: my_gc_assm_name
  report_usage: boolean
  submol: File
  ignore_all_errors:
        type: boolean?
  no_internet:
    type: boolean?
  make_uuid:
    type: boolean?
    default: true
  uuid_in:
    type: File?

What I confused is: Are the blast_hits_cache_data and uuid_in required for the input? I can't find them in the input directory.

I'd appreciate it if anyone could help

thanks

[azat-badretdin]

Apologies for not noticing your earlier comment. Good news: you do not have anymore input file problems.

Bad news. The first mention of permanentFail indicates that there is some problem with the system, most likely:

/panfs/pan1.be-md.ncbi.nlm.nih.gov/gpipe/bacterial_pipeline/system/2022-04-14.build6021/arch/x86_64/bin/sparclbl_mp.sh: line 384: 520393 Segmentation fault      (core dumped) "${rpsblast_bin}" "${BLASTCMD[@]-}" "${blastargs[@]-}" -query "${single_seqfile}" -out "${sglasn}"
+ excode=139
+ [[ excode -ne 0 ]]
+ echo 'Rpsblast failed with return code 139, abort ...'
Rpsblast failed with return code 139, abort ...

I would recommend following options: try to modify cpus parameter: very low value (8), no parameter and see if this helps.

If the problem persists please open a new ticket for this problem, so it will be easier for readers to focus on their specific problem. Thanks!