#288 Ghost SGE file and partition

[Ludovic Mallet]

When running pacBiotoCA 8.2 beta precompiled:
Message of complaint about sge.out, while the job was launched without the grid.

extract of the used spec file:

grid info

useGrid = 0
scriptOnGrid = 0
frgCorrOnGrid = 0
ovlCorrOnGrid = 0

ls: cannot access /global/group/home/lmall_01/projects/Bt/assembly/247/v_9//tempBt247_Subreads/sge.out: No such file or directory
* Correction Summary *
Illegal division by zero at /opt/global/bin/pacBioToCA line 2720.

The second error is more cryptic

the frag was created, but other files remain empty.

0 Nov  6 18:15 Bt247_Subreads.fasta
0 Nov  6 18:15 Bt247_Subreads.qual
0 Nov  6 18:15 Bt247_Subreads.fastq

645 Nov 6 18:15 Bt247_Subreads.frg
0 Nov 6 18:15 Bt247_Subreads.log

Here is the out file:
Running partition 000001 with options -h 6573397-7194731 -r 1-6573396 --hashstrings 621335 --hashdatalen 236621579 start 0 end 6573396 total 6573396 zero job 0 and stride 1
Blasr completed.
SamToCA conversion completed.
Total split bases is 0 vs 236000244 so ratio is 0
Total corrected bp (before consensus): 0 (0.000).

[Sergey Koren]

Hi,

The message regarding missing files is OK. The script has generic post-processing to collect all output whether it is run on SGE or not. When it is on SGE, it will collate all the individual output files but when it is run on a local system the output is just sent to stdout/err.

The second message seems to indicate no data could be corrected. I've not seen that error before. Have you been able to run the sample data from the wiki successfully? Can you provide the full output of the pipeline up to the point it fails?

[jeff]

Hey there, I actually just ran into this same issue (identical error output), have you found a resolution to the error? Thanks

[Sergey Koren]

The SGE error is fine and can be safely ignored as long as your fasta/fastq files are not empty.

[Sergey Koren]

Closing. Error was due to a BLASR crash which as being run as part of runPartition. Rebuild of BLASR fixed the error.