Opened 11 years ago

Closed 11 years ago

#288 closed error (fixed)

slab hadam3 job not ticking over more than 1 month

Reported by: agt Owned by: jeff
Component: UM Model Keywords: slab, crun, 1 month only
Cc: Platform:
UM Version: 4.5

Description

Hi all,

I have a strange problem with the slab model of vn4.5.

This seems to be happening since the hector changes but I can't be sure.

I was originally trying to run with umcet and the same problem was occuring, so I have stepped back to a standard job and the problem still happens.

Job is xedaa (compile/reconf and stop) and then xedae (run from existing). I am operating separately as that is necessary for UMCET. The xedaa part seems to create an executable just fine (see ~agt/um/umui_out/xedaa000.xedaa.d09156.t174948.leave) although note there are lots of warnings about "mc" in there).

Job xedae then runs from exsiting using the reconf. dump. In this job I reqeust 3600s on 4x2 processors. At this rate I normally get around 1month in 15 minutes. From the timestamps in the output directory ~agt/work/um/xeda you can see that xedae started at about 18:30 and the april months files finished at around 18:45. The xedaea.pai0apr file (daily means) seems to have "sensible" looking output at day 30. However, then the jobs does nothing: no monthly .da for the start of May. It just sits there, uses up the rest of the hour requested, and then at 19:30 the xedae000.xedae.d09156.t182928.leave tells me that walltime is exceeded.

any help is appreciated, cheers, Andy

Change History (6)

comment:1 Changed 11 years ago by jeff

  • Owner changed from um_support to jeff
  • Status changed from new to accepted

Hi Andy

I'm not sure whats wrong here, the easiest thing for me to do is to try and run it myself. To do this I need read permission on your files under /work/n02/n02/agt.

Jeff.

comment:2 Changed 11 years ago by jeff

Hi Andy

Your run seems to be hanging when it is trying to compress slab stash field 40,226 (GRID BOX AREAS), I'm not sure why. If you turn off packing for the .pf file then it works, i.e. in umui panel

Sub-Model Independent → Post Processing → Intialisation and processing of mean & standard PP files

Change the packing profile for PP5/PF/65 from 5 to 0. Also is this field time varying? It doesn't look very interesting.

Jeff.

comment:3 follow-up: Changed 11 years ago by agt

Jeff,

I'll try first with leaving the packing as it is but removing that pointless diagnostic from stash.

thanks,

Andy

comment:4 in reply to: ↑ 3 Changed 11 years ago by jeff

Replying to agt:

Jeff,

I'll try first with leaving the packing as it is but removing that pointless diagnostic from stash.

Did this work?

Jeff.

comment:5 Changed 11 years ago by agt

Jeff,

yes thanks, removing the grid areas diagnostic from stash solved the problem. Therefore I did not need to alter the packing,

please close the ticket,

thanks,

Andy

comment:6 Changed 11 years ago by jeff

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