Wiki

Clone wiki

bnpy-dev / BrownCSGridFAQ

Grid Basics

Where do I see a list of machines on the Brown grid?

http://cs.brown.edu/about/system/services/hpc/grid/

How many jobs can I run at one time?

Generally, it's good manners to keep the number of jobs you submit to less than ~200 "short" (<1 hr) jobs, and less than ~50 "long" (<24hr) jobs.

If the grid is busier than usual, you may wish to run fewer jobs, while if the grid is not used at all, you can probably run a few more. Remember that you can always use qstat -u '*' to see all active jobs on the grid, so you can get a sense of which users are active and who you should be worried about competing with.

Troubleshooting

I can't use ssh/qlogin... I get RSA key fingerprint errors... do I want to continue?

Probably, you need to reauthorize your credentials. Open a terminal and run the kinit command.

$ kinit
You'll be prompted for your LDAP password. Please enter it. Hopefully this fixes the problem. Otherwise, try again later.

Who do I talk to about problems with the grid?

Email problem@cs.brown.edu. Usually John Bazik is the one to talk to.

Updated