Re: [Condor-users] Condor for Fedora Core 2 with 2.6 Kernel ??

[Erik Paulson <epaulson@xxxxxxxxxxx>]

On Thu, Jan 06, 2005 at 09:11:35PM -0800, John Wheez wrote:
> Erik Paulson wrote:
> 
> >I'd like to see the StartLog from the FC2 machine, as soon as the 
> >condor_startd starts up
> >
> >Also, the CollectorLog from the central manager. 
> >
> >-Erik
> > 
> >
> Hi Erik,
> 
> I've done more R&D and my Redhat Master server can control the Fedora 
> Core Two machine via condor_off and condor_on etc...The fedora core 2 
> machine still doesn't show up in teh output of conndor_status at all.
> 

Your collector log is not nearly long enough - for one (and maybe this 
is just because you don't have your clocks synchronized), but your 
collectorlog is from 3 minutes before you turned on your startd.

Please put on a webpage somewhere, or send to condor-admin, the 
_full_ logfile. It'd be best if you included machine names and 
IP addresses of the machines involved, and the best experiment to do
is to:

1. Turn Condor off on both the central manager and FC2 machine.
2. Turn Condor back on on the central manager, wait 30 seconds or so
3. Turn condor back on on the FC2 machine, and give it 30 seconds or
so. 

Then do a 'condor_status -l' and a 'condor_status -master -l', and a 
'condor_status -direct <hostname_of_fc2_machine -l'

I doubt there's anything wrong with yoru FC2 startd - if it couldn't
figure out how much memory or CPUs there were in the machine it would
have complained - it looks very much like it calculated it right.

Are you sure CONDOR_HOST is pointing to the right place on the FC2
machine? When you run condor_status on that machine, what shows up?

-Erik

> 
> 
> Here is the output of the Fedora Core 2 Machines StartD ..the output of 
> the collector is below
> 
> 1/6 21:03:35 ******************************************************
> 1/6 21:03:35 ** condor_startd (CONDOR_STARTD) STARTING UP
> 1/6 21:03:35 ** /opt/condor-6.7.3/sbin/condor_startd
> 1/6 21:03:35 ** $CondorVersion: 6.7.3 Dec 28 2004 $
> 1/6 21:03:35 ** $CondorPlatform: I386-LINUX_RH9 $
> 1/6 21:03:35 ** PID = 2809
> 1/6 21:03:35 ******************************************************
> 1/6 21:03:35 Using config file: /opt/condor-6.7.3/etc/condor_config
> 1/6 21:03:35 Using local config files: 
> /opt/condor-6.7.3/local.thunder/condor_config.local
> 1/6 21:03:35 DaemonCore: Command Socket at <192.168.0.3:32849>
> 1/6 21:03:36 vm1: New machine resource allocated
> 1/6 21:03:36 vm2: New machine resource allocated
> 1/6 21:03:36 vm3: New machine resource allocated
> 1/6 21:03:36 vm4: New machine resource allocated
> 1/6 21:03:36 About to run initial benchmarks.
> 1/6 21:03:40 Completed initial benchmarks.
> 1/6 21:03:40 vm2: State change: IS_OWNER is false
> 1/6 21:03:40 vm2: Changing state: Owner -> Unclaimed
> 1/6 21:03:40 vm3: State change: IS_OWNER is false
> 1/6 21:03:40 vm3: Changing state: Owner -> Unclaimed
> 1/6 21:03:40 vm4: State change: IS_OWNER is false
> 1/6 21:03:40 vm4: Changing state: Owner -> Unclaimed
> 
> 
> Collector log on Linux Redhat WS3 Host
> 
> 1/6 21:00:50 Got INVALIDATE_SCHEDD_ADS
> 1/6 21:00:50            **** Removing stale ad: "< thunder.atomfx.com , 
> 192.168.0.3 >"
> 1/6 21:00:50 (Invalidated 1 ads)
> 1/6 21:00:50 Got INVALIDATE_STARTD_ADS
> 1/6 21:00:50 (Invalidated 0 ads)
> 1/6 21:00:50 (Invalidated 0 ads)
> 1/6 21:00:50 Got INVALIDATE_STARTD_ADS
> 1/6 21:00:50 (Invalidated 0 ads)
> 1/6 21:00:50 (Invalidated 0 ads)
> 1/6 21:00:50 Got INVALIDATE_STARTD_ADS
> 1/6 21:00:50 (Invalidated 0 ads)
> 1/6 21:00:50 (Invalidated 0 ads)
> 1/6 21:00:50 Got INVALIDATE_STARTD_ADS
> 1/6 21:00:50 (Invalidated 0 ads)
> 1/6 21:00:50 (Invalidated 0 ads)
> 1/6 21:00:50 ScheddAd     : Inserting ** "< thunder.atomfx.com , 
> 192.168.0.3 >"
> 
> On Thu, Jan 06, 2005 at 10:20:22AM -0800, John Wheez wrote:
> 
> >Hi Eric,
> >
> >I tried condor 6.7.3 Fedora Core 1 RPM on my Fedora Core 2 system and it 
> >still doesn't work correctly.
> >The problem is that teh classadds from the Fedora Core 2 system are 
> >never accepted by the condor pool host.
> >
> >I do not see any errors in the logs on my host either.
> >
> >But condor_status never shows the Fedora Core 2 system in it's output.
> >
> > 
> >