Stats

Locked Locked
Replies 6
Subscribers 126
Views 17745
Members are here 0

This discussion has been locked.

You can no longer post new replies to this discussion. If you have a question you can start a new discussion

LSF jobs never finishing

Hello,

I've been using LSF to launch ADEXL jobs for some time now without issue but lately I've been having trouble with LSF jobs launched from ADEXL not finishing. The individual APS jobs finish and I get results but I have to manually go kill a ton of jobs from the LSF queues because ADEXL doesn't seem to be cleaning up after itself. Has anyone seen this before?

Here is some more detail. On the "Job Policy Setup" I have
Start Timeout = 120
Configure Timeout=300
LingerTime=60

Also, I'm using the "Command" Distribution Method (not LBS). I have to use the command because I need to reserve multiple processors and to my knowledge this isn't possible with LBS. The specific command I use is

bsub -q medium -n 4 -R 'span[hosts=1]'

Any thoughts?

Thanks

Andrew Beckett over 13 years ago

What version (subversion) are you using? That shouldn't happen...

Regards,

Andrew
Cancel
Vote Up 0 Vote Down

Cancel
BradW over 13 years ago

Right now I’m running 6.1.5-64b.500.5
Cancel
Vote Up 0 Vote Down

Cancel
Andrew Beckett over 13 years ago

Can you log a service request with customer support? This needs investigation, I think.

Andrew
Cancel
Vote Up 0 Vote Down

Cancel
BradW over 13 years ago

Will do. Thanks for the quick response.
Cancel
Vote Up 0 Vote Down

Cancel
mhowell67 over 13 years ago

Hi Andrew, I've also seen this issue and I can duplicate it reliably. SR 42657658
It only happens when using "command" distribution method. The various options on bsub don't seem to matter. I've done single runs with no multi-threading, corners with and without mt, and monte carlo and everytime the JOBID must be killed manually.
Using LBS distribution method does not exhibit this issue nor have I seen this with ADEL.
I've observed this behavior with several versions of 6.1.4 and 6.1.5.500.6.
Almost forgot to mention. This is not a case of the "lingering". I can close all data, shutdown ADEXL and close my virutoso session and the JOBID will still be listed in LSF.
Cancel
Vote Up 0 Vote Down

Cancel
Andrew Beckett over 13 years ago

Thanks - I'll keep an eye on the CCR that was filed with R&D for this.
Andrew.
Cancel
Vote Up 0 Vote Down

Cancel

Community Guidelines

The Cadence Design Communities support Cadence users and technologists interacting to exchange ideas, news, technical information, and best practices to solve problems and get the most from Cadence technology. The community is open to everyone, and to provide the most value, we require participants to follow our Community Guidelines that facilitate a quality exchange of ideas and information. By accessing, contributing, using or downloading any materials from the site, you agree to be bound by the full Community Guidelines.