[Mauiusers] Maui unresponsive while writing strange lines unto log file

David B Jackson jacksond at clusterresources.com
Fri Dec 14 22:58:10 MST 2007


Josh,

  Can you roll this patch into the latest release of Maui?

Thanks,
Dave

> It is debug printing that happens when it detects corruption in
> the reservation table. (Why that happens is another issue...)
>
> I've attached a patch that we use to turn off this printing unless
> the log level is turned up.
>
> Tom
>
>
> Manuel Reiter wrote:
>> Hi,
>>
>> I'm running maui 3.2.6p14 and torque 2.0.0p8 on a ~250 node Opteron
>> cluster. While scheduling works fine, maui is often unresponsive to
>> commands like showq, showres and so on. While this is the case, maui
>> seems to write many lines of the form
>>
>> 12/13 13:45:23 INFO:     R1[109]  S: 1197730705  E: 1197730718  T:  170
>> N: 92
>>
>> into the log file, although I have
>>
>> LOGLEVEL              0
>>
>> and, experimentally, even
>>
>> LOGFACILITY             fLL
>>
>> in my maui config file.
>>
>> Today alone, maui has written about 200.000 of these lines in about 5
>> hours. The pattern is that the index after R1 goes from 0 to 254,
>> followed by two lines like the above but with R1[n] replaced by R2[0]
>> and R2[1] and then things start over. Between these bursts, I have
>> hours when none of these lines appear in the log and maui is quite
>> responsive.
>>
>> Can anybody tell me what these lines actually mean and why maui is
>> spitting out so many of them? Or provide any other insight into my
>> problem os unresponsiveness? I have put
>>
>> RMPOLLINTERVAL        300
>> NODEPOLLFREQUENCY       20
>> JOBAGGREGATIONTIME 60
>>
>> in the maui config in the hopes that this would make things better,  but
>> this didn't change things.
>>
>> On another cluster I run (same maui, torque 2.0.0p4) no similar lines
>> appear in the maui logs at all.
>>
>> Any help would be greatly appreciated.
>>
>> Thanks and best regards,
>>
>>   Manuel
>>
>> ------------------------------------------------------------------------
>> ------
>> Manuel Reiter                      |         reiter at th.physik.uni-
>> frankfurt.de
>> Center for Scientific Computing    |
>> J.W.Goethe Universität             |
>> D-60054 Frankfurt am Main          |
>> Germany                            |
>> _______________________________________________
>> mauiusers mailing list
>> mauiusers at supercluster.org
>> http://www.supercluster.org/mailman/listinfo/mauiusers
>>
>
> _______________________________________________
> mauiusers mailing list
> mauiusers at supercluster.org
> http://www.supercluster.org/mailman/listinfo/mauiusers
>



More information about the mauiusers mailing list