Job submission error (CRAY)

[dan2012]

Hi,

I am attempting to start one of my HadGem? simulations on the CRAY for the first time, and receive this error when submitting on the UMUI:

ERROR: can't read "pwstatus(xcml00,dapart)": no such element in array while attempting to access account dapart on host xcml00.

(Please respond to the cc address also, dan.partridge at aces.su.se, and my Oxford e-mail is currently not receiving external mail).

Any help on this would be greatly appreciated.
Many thanks,
Daniel

[grenville]

Daniel

Please just try again - I believe this happens to new Cray MONSooN users one time only.

Grenville

[dan2012]

Thanks Grenville,
I tried again and this time receive a new error:
ERROR: lander.monsoon-metoffice.co.uk: Permission denied while attempting to access account dapart on host xcml00.
I can log into monsoon and xcm fine through terminal however.
Best,
Dan

[ros]

Hi Dan,

Have you set up your ssh-agent as detailed here: http://cms.ncas.ac.uk/wiki/MonsoonSshAgent

Cheers,
Ros.

[grenville]

Dan

Not sure, do you normally use /projects/ukca-meto/mdalvi as the target machine route directory - this is usually your space.

Grenville

[dan2012]

Dear Grenville,
Thanks for this, I had indeed missed changing one path in the UMUI that was causing this problem.
The job did now successfully submit, however, there is nothing in /output on the XCM.
On the new cray system, where does one check if a job crashes once it has submitted successfully?
Many thanks,
Dan

[ros]

Hi Dan,

The job didn't submit successfully, you are trying to use an invalid project/account code: ukca-meto (See /projects/ukca-ox/dapart/xlyoa/umrecon/ext.out). I guess you need to change the account group to ukca-ox in the General details window.

Once the job has submitted successfully the .leave files will appear in ~/output.

Cheers,
Ros.