[GE users] Strange qstat -g c -ext Vs qstat -qs E

Reuti reuti at staff.uni-marburg.de
Fri Aug 4 17:38:53 BST 2006


Am 04.08.2006 um 17:39 schrieb Amit H Kumar:

>

Now I'm confused:

>
> Reuti <reuti at staff.uni-marburg.de> wrote on 08/04/2006 10:59:53 AM:
>
>> Hi,
>>
>> Am 04.08.2006 um 15:57 schrieb Amit H Kumar:
>>
>>>
>>> Hi SGE,
>>>
>>> We noticed a strange thing.
>>>
>>> When we "qstat -q c -ext"
>>> We see that one of the queues has a node in error state.
>>>

was this a typo and you meant -g, not -q? It looked to me like your  
queue was named c ;-), hence it would be an job in error state listed  
at the end.

>>> But when we do "qstat -qs E": It doesn't show anything.
>>>
>>> Though I can see this using QMON. and it shows that the error was
>>> due to a
>>> job failure.
>>
>> is it possible, that not the queue, but the job is in error state?
>> There are two qmod options to clear it:
>
> Okay. I see the following as a result of:=> qstat -explain E
>       queue large.q marked QERROR as result of job 10371's failure  
> at host
> mycluster-0-48.local
>

Okay. This is also shown in "qstat -f" as status E - right?

>> From the above output can I be certain that the error was due to a  
>> Job's
> failure only.
> And if yes, is this the only way to check and tell if it was an  
> error due
> to job failure.
>
> Does it also mean that: "qstat -qs E" shows only errors due to que  
> and not
> due to job failure ?

The -qs option is a filter(*), not a selection, for the conventional  
output. Please try:

qstat -f -qs E

(at least for me it's working this way).

-- Reuti


(*) Is the man page wrong, as it states to behave like -f, or the  
implementation of -qs - any comments?

PS: There is issue http://gridengine.sunsource.net/issues/ 
show_bug.cgi?id=2073 for u4, but it seems already to be fixed in u7 -  
any comments?


>
>>
>>         -cj    Clears the error state of the specified jobs(s).
>>
>>         -cq    Clears the error state of the specified queue(s).
>>
> Thank you Reuti for you help and feedback
> -AK
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
> For additional commands, e-mail: users-help at gridengine.sunsource.net

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net




More information about the gridengine-users mailing list