Jobs stuck in Eqw or hqw states

[wyim-pgl]

Jobs stuck in Eqw or hqw states

[wyim-pgl]

It looks like your computing nodes cannot access password system.

Usually, you don't need to do anything special authenticate for SGE job submission, because if you already login to login node, SGE collect the your information from login node.

However, if the compute nodes are also under kerberos authenticate, then you will need to
enable security when you compile SGE, that is, add the -kerberos flag (and make sure you have the kerberos development libraries on the machine where you compile the source, etc)

Also these errors appear either consistently or inconsistently and are no longer correlated with the use of OpenDirectory (problems appear with local user accounts as well), NFS, case-insensitive filesystems or any other system settings.

More information and background on this issue can be found via these links

https://linuxfollies.blogspot.com/2014/05/sssd-setup-in-bright-cluster.html

Can you do qrsh or qsub for your job?

simple tests are below

qrsh hostname

 qsub -b y -t 1-40 hostname

or

login as 'henry' instead of Henry.

If you have hetero HPC system with OSX and linux. It should be a problem.

If you cannot find anything, I hope you contact to your admin. Because this is not the DCBLAST problem.

[Amy-T]

Thanks for the suggestions. We will give it a try.