Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

I want to mark a node (or set of nodes) as "offline" in the sense that I want Sun Grid Engine to stop assigning new work to them. This would be for some kind of maintenance work on the nodes themselves. The nodes should finish whatever work they've been assigned, and then just go into some kind of idle ("offline") state. I've been hunting through the qconf documentation, but I can't find this use case in any howto.

share|improve this question

1 Answer 1

up vote 3 down vote accepted

Searching about has led me to the qmod utility. I've done a simple test of

qmod -d QUEUENAME.q@MACHINENAME

and this seems to be working, although I haven't actually tried it with jobs running. The qstat output is changed to indicate that the node is disabled - a "d" flag appears.

qmod -e QUEUENAME.q@MACHINENAME

will enable the machine again.

On our cluster, the machines are named worker-##-## where the two numbers are rack number and rank number. We only run one master queue, called "all.q". And the machines in our cluster get listed with a ".local" suffix in the qstat output. So the above command ends up being

qmod -d all.q@worker-9-9.local

to take the machine at rack 9, rank 9 out of the queueing rotation.

share|improve this answer
    
I've been using this command a bit today to do cluster maintenance. I believe it's the right thing. –  Rick Reynolds Sep 22 '09 at 16:23

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.