[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-ha
Subject:    Re: [Linux-HA] High available job scheduler
From:       Alan Robertson <alanr () unix ! sh>
Date:       2004-01-22 2:00:20
Message-ID: 400F2EB4.5010303 () unix ! sh
[Download RAW message or body]

Karl Kopper wrote:
>>I am working on on a project which involves high available job scheduler
> 
> at failover state at the master node. which is the best scheduler to start
> with. i have been >looking at SGE which has some failover functionality i
> need some advice on this
> 
>>Thanks in Advance
>>varma
>>
> 
> 
> One solution is to use the Ganglia package and the gexec (and authd)
> programs. If you only need to run jobs on one node a command like:
> 
> gexec -n 1 command
> 
> will pick the best (least loaded) node that has Ganglia (gmond) installed on
> it. It is not the most sophisticated batch scheduling system, but it is
> pretty simple to set up and use.
> 
> Using Ganglia with gexec has the added benefit that you get very nice
> performance/load graphical displays (with the Ganglia web frontend) of your
> cluster or of your cluster of clusters.
> 
> The OSCAR project uses a different batch job scheduler--you might also want
> to check out OSCAR as well.

Is the ganglia scheduler highly-available?  That is, can it work with 
something like heartbeat to provide highly available scheduling services.

It should be able to do that if it writes status information to a disk file, 
and it can be restarted from that disk file, AND if its data structures on 
disk can withstand a crash.

Do you know if it does that?


-- 
     Alan Robertson <alanr@unix.sh>

"Openness is the foundation and preservative of friendship...  Let me claim 
from you at all times your undisguised opinions." - William Wilberforce

_______________________________________________
Linux-HA mailing list
Linux-HA@lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha
[prev in list] [next in list] [prev in thread] [next in thread]