[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Condor-users] unhandled DEFAULT_PRIO_FACTOR ?



Hi Max
 
HAD was introduced in 6.7.6 development series.
In out lab in DSL we have some intermediate NOT official version of 
6.6.9 with HADs.

One good idea would be to upgrade Condor in DSL to an official 6.7.7.

You are right that :
> I think it is possible that in HAD situation, when killing 1 manager 
> (out of 3 ) will
> pass the ball to the left 2, which probably mess some things.
 
But recall that HAD have a meaning only when Neg machine failed, and in
any other situation it is irrelevant. If you have updated config for
only one Neg and restarted it, there is some chance that another Neg
will be started instead and if that one was not configured right (with
DEFAULT_PRIO_FACTOR such that you want) than it would keep an old
DEFAULT_PRIO_FACTOR.
 
On the other hand, if HADs in DSL are configured to USE_PRIMARY_HAD
(and as far as I remember this was the sitiuation last time I checked),
than it means that the first primary Neg will be reelected when
restarted and in such case the problem has nothing to do with HADs.

Regards,
Gabi

-----Original Message-----
From: condor-users-bounces@xxxxxxxxxxx
[mailto:condor-users-bounces@xxxxxxxxxxx] On Behalf Of Maxim Kovgan
Sent: Thursday, June 02, 2005 7:30 PM
To: Erik Paulson
Cc: Condor-Users Mail List
Subject: Re: [Condor-users] unhandled DEFAULT_PRIO_FACTOR ?


Hello, Erik.

> Check the MasterLog on your central manager. My guess is that your
Nothing of this sort there: 
For the matter of doubt removal I restarted the central manager..

> condor_reconfig was denied and so the running daemons didn't get the 
> change. The MasterLog should give you enough information about what 
> went wrong, or just send a SIGHUP to the condor_collector

Our version of condor might be not exactly straight plain vanilla 6.6.9
I've noticed "HAD" in DAEMON_LIST in condor_config. when was HAD
introduced into the mainstream ?  

It can be the issue of HAD changes handling:
I think it is possible that in HAD situtation, when killing 1 manager (
out of 3 ) will 
pass the ball to the left 2, which probably mess some things.

Thanks for the input,
Max.

> -Erik

_______________________________________________
Condor-users mailing list
Condor-users@xxxxxxxxxxx
https://lists.cs.wisc.edu/mailman/listinfo/condor-users