StarCluster - Mailing List Archive

Re: addnode question

From: Justin Riley <no email>
Date: Thu, 23 Feb 2012 18:47:16 -0500

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Also, are you using t1.micro instances by chance? If so you'll want to
check the load and memory usage on the master node and the new node
after starting:

$ uptime
$ free -m

The micro instances run out of resources very quickly which could cause
your jobs to stay queued until the load averages go down.

~Justin

On 2/23/12 5:56 PM, Justin Riley wrote:
>
> Assuming the node is successfully added to SGE new jobs should start
> running on the new host. I need more details to help you with this. Does
> running 'qhost' on the master node show the new node after running
> addnode for example? Also how did you submit your jobs? What are your
> jobs slot requirements?
>
> If you could attach the output of 'qhost' and 'qstat' after queueing
> jobs and running 'addnode' that would be helpful in figuring out what
> the issue is.
>
> ~Justin
>
> On 2/23/12 1:56 PM, Robert Yu wrote:
> > Hi,
>
> > I have many jobs queued up. I then run "addnode" but these jobs don't
> > seem to take advantage of the new resource. Is there something I need
> > to do to inform sge there are new resources? Maybe delete and
> > resubmit? Or maybe restart the sge daemons?
>
> > Thanks.
>
>
>
> _______________________________________________
> StarCluster mailing list
> StarCluster_at_mit.edu
> http://mailman.mit.edu/mailman/listinfo/starcluster

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk9G0AQACgkQ4llAkMfDcrmEIACfZ5nInEE/4t2Sat92aepabLG0
KK8AnjfXVqMarCvw3WEZAkcGaIXouh5j
=I4We
-----END PGP SIGNATURE-----
Received on Thu Feb 23 2012 - 18:47:18 EST
This archive was generated by hypermail 2.3.0.

Search:

Sort all by:

Date

Month

Thread

Author

Subject