Hi,
I've got all condor on 8.2.8 or 8.3.6. Updgrade was just stop condor,
'yum update', start condor.
raul
On 02/07/15 12:23, Andrew Lahiff wrote:
> Hi Gang,
>
> As far as I remember it was a simple upgrade. You shouldn't even need to restart condor services - condor_master will do this for you (and even drain worker nodes first before restarting if you have it configured to do so).
>
> Regards,
> Andrew.
>
> ________________________________________
> From: Testbed Support for GridPP member institutes [[log in to unmask]] on behalf of qing [[log in to unmask]]
> Sent: Thursday, July 02, 2015 11:42 AM
> To: [log in to unmask]
> Subject: Re: Condor held jobs
>
> Hi, Andrew:
>
> At Glasgow we also see such errors with condor 8.2.2, but it happens
> not very frequently.
>
> We will migrate to 8.2.7, anything particular needs to be concerned?
> Or just a plain upgrade, e.g. yum upgrade condor and then restart condor
> service is enough.
>
> Cheers,Gang
>
> On 02/07/2015 11:34, Andrew Lahiff wrote:
>> Hi Kashif,
>>
>> What version of condor are you using? This problem should now be fixed (at least it is for us with 8.2.7).
>>
>> Regards,
>> Andrew.
>>
>> ________________________________
>> From: Testbed Support for GridPP member institutes [[log in to unmask]] on behalf of Kashif Mohammad [[log in to unmask]]
>> Sent: Thursday, July 02, 2015 11:22 AM
>> To: [log in to unmask]
>> Subject: Condor held jobs
>>
>>
>> Hi
>>
>> I am investigating an issue with Condor held jobs. Some of the jobs stay in Held state for long time. The jobs are failing with Close_Pipe error and tracing back to the worker node, it shows this error
>>
>> ERROR "Close_Pipe error" at line 2089 in file /slots/02/dir_35085/userdir/src/condor_daemon_core.V6/daemon_core.cpp
>>
>> Andrew Lahiff mentioned same problem on HTCondor-users list but it seems that it didn’t reach on any conclusion.
>>
>> Andrew, did you manage to fix this issue? Any one else has seen this?
>>
>> Thanks
>>
>> Kashif
|