Opened 3 years ago

Closed 3 years ago

#1945 closed help (answered)

job limit

Reported by: simon.tett Owned by: ros
Component: ARCHER Keywords:
Cc: Platform: ARCHER
UM Version:

Description

Hi,

I am doing data conversion of field files to netcdf by submitting jobs to the serial Q. It has all been running smoothly till now but I now have an ensemble of jobs.. So lots of conversion jobs. I get an error:

qsserver: Sat Aug 13 08:43:06 BST 2016: xmvpaa.ps1915son ARCHIVE PPNOCHART
qsub: would exceed complex's per-user limit
Generated Conversion Script /work/n02/n02/stett2/tmp/tmp.mom5.5772/xmvpaa.ps1915son.ff_arch_31308.txt
qscasedisp: return code after calling qsarcher_convnc RCARC=38

Two of my jobs failed as a consequence.. Can my limit on the number of jobs be doubled?

thanks

Simon

Change History (4)

comment:1 Changed 3 years ago by ros

  • Owner changed from um_support to ros
  • Status changed from new to accepted
  • UM Version <select version> deleted

Hi Simon,

The number of jobs that you can submit to each queue is a hard-limit set by ARCHER and is not changeable. The maximum jobs you can submit to the serial queue is 12.

Regards,
Ros.

comment:2 Changed 3 years ago by simon.tett

A sort of update to this. I think what happened was the serial Q got blocked by a long lasting job. So all my jobs piled up and then the model crashed. This per say is not a bad outcome — if the model just kept going it would fill up disk space and crash in a more uncontrolled way. Maybe it illustrates a problem with my approach of using jobs to convert data rather than ssh. However, ssh is very unreliable and I found that failed very frequently…

So does the serial Q only allow one job to run at a time?

Simon

comment:3 Changed 3 years ago by ros

Hi Simon,

You can submit a maximum of 12 jobs to the serial queue per user only 6 of which can be running at any time.

A full list of limits is available on the archer website:
http://www.archer.ac.uk/documentation/user-guide/batch.php#sec-5.8

Cheers,
Ros.

comment:4 Changed 3 years ago by ros

  • Platform set to ARCHER
  • Resolution set to answered
  • Status changed from accepted to closed
Note: See TracTickets for help on using tickets.