[GE users] Upgrading/migrating from 5.3 to 6.0

Shane Canon canon at nersc.gov
Thu Apr 7 19:21:33 BST 2005

Hash: SHA1


We have recently seen poor responsiveness from the qmaster in our SGE
5.3 system.  This occurs with around 10k jobs, 200 nodes, and several
dozen active users.  I suspect our queues (which are slightly complex)
may be contributing as well.  The consensus on the list seems to be that
we should upgrade to 6.0.  So, I am in the process of testing how the
migration will go.  I have a few questions.

1) How much will sticking to classic spooling hurt us from a performance
stand point?  I prefer to stick with classic spooling for now, since we
still want shadowing and I am not completely comfortable with running
the rpc based DB server.  We will still see a strong benefit with 6.0
without using berkeleydb?

2) Is there a way to convert pending jobs from 5.3 to 6.0?

A few other unrelated questions,...

Are there any plans to add advanced reservations (not resource
reservation) to SGE?

Has anyone looked at Maui integration with SGE 6.0?  The last messages I
saw on this said that the effort had stalled.

How is the usage scaling in the exec host config used?  Is this used in
the share tree calculation?  Would this be the best place to correct for
differences in processor power between heterogeneous nodes?

Well, that's probably enough for now.


- --Shane
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org


To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net

More information about the gridengine-users mailing list