[GE users] Frontend as exec host
dasf2008 at gmail.com
Mon Apr 12 19:09:15 BST 2010
[ The following text is in the "iso-8859-1" character set. ]
[ Your display is set for the "ISO-8859-10" character set. ]
[ Some special characters may be displayed incorrectly. ]
You guys are great! Thanks, Fernando. Your recipe worked out just fine!
On Mon, Apr 12, 2010 at 11:33 AM, rozenblit <rozenblit at gmail.com<mailto:rozenblit at gmail.com>> wrote:
Is this a Rocks 5.2 system? In a small cluster, there's no problem
using the frontend as execnode (I just limit it to 6 of the 8 cores).
This seems like a problem I had some /time ago. You could have a look
at this: http://gridengine.sunsource.net/howto/multi_intrfcs.html
Or, if you're using rocks, you're probably going to do this:
1) Create /opt/gridengine/default/common/host_aliases with the following:
2) Edit /opt/gridengine/default/common/act_qmaster to be [frontend].local
3) Remove the last line from /etc/init.d/sgemaster.[frontend]
("/bin/hostname --fqdn > /opt/gridengine/default/common/act_qmaster")
On Thu, Apr 8, 2010 at 7:52 PM, dasf <dasf2008 at gmail.com<mailto:dasf2008 at gmail.com>> wrote:
> Ladies and Gentlemen,
> I am trying to add the frontend as an execution host but I am having some weird problems. I added as many suggested on the web, i.e., running ./exec_host at the frontend and accepting all questions. Everything looked fine, as below. The cluster.local was added, I see the load, slots available, etc...
> But as soon as I add jobs to fill up the entire cluster, problems appear. First, no jobs start to run on the frontend. Second, the load_avg changes to -NA-.
> Any clue on how to fix it?
To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net<mailto:users-unsubscribe at gridengine.sunsource.net>].
More information about the gridengine-users