[GE users] weird "orders queue version is not uptodate" messages in qmaster log

Xavier MACHENAUD xavier.machenaud at st.com
Fri Mar 4 08:19:06 GMT 2005


Stephan,

Just one more question : Do you know how to recover from the blocked 
state (when queued jobs are not started anymore)?

Xavier

Terry Lalonde wrote:

>I too am seeing this with u1.  
>
>I do not plan to upgrade soon.
>
>  
>
>>-----Original Message-----
>>From: Stephan Grell - Sun Germany - SSG - Software Engineer
>>[mailto:stephan.grell at sun.com]
>>Sent: Thursday, March 03, 2005 5:03 AM
>>To: users at gridengine.sunsource.net
>>Subject: Re: [GE users] weird "orders queue version is not uptodate"
>>messages in qmaster log
>>
>>
>>
>>Xavier MACHENAUD wrote:
>>
>>    
>>
>>>FYI, I also got the messages in u1 and reach to point where no jobs
>>>      
>>>
>were
>  
>
>>>scheduled anymore.
>>>
>>>I upgraded to u3 4 days ago. I still have the messages but, so far,
>>>didn't reach the blocking state.
>>>But I'me not very confident about not being blocked again :-(
>>>
>>>      
>>>
>>Well, if you are brave, you could try the maintrunk. :-)
>>This problem should be fixed. Would be nice to have an external test
>>    
>>
>for
>  
>
>>it.
>>
>>Stephan
>>
>>    
>>
>>>Xavier
>>>
>>>Stephan Grell - Sun Germany - SSG - Software Engineer wrote:
>>>
>>>
>>>
>>>      
>>>
>>>>Hi Terry,
>>>>
>>>>its a way to ensure, that the scheduler uses the latest data for
>>>>its scheduling decision. Every object has a version number. When
>>>>the scheduler generates updates for an object (example: job start),
>>>>it includes the version number into that order to inform the qmaster
>>>>about the basis of its decision.
>>>>
>>>>Unfortunately, there is a bug in u3, which delays the delivery of
>>>>        
>>>>
>update
>  
>
>>>>events. Therefore, the scheduler is not working on the most recent
>>>>data, and the qmaster logs the error messages you have noticed.
>>>>
>>>>This can cause:
>>>>- losing usage in the sharetree.
>>>>- trying to start a job twice
>>>>- ignoring job start orders
>>>>
>>>>The likely hod of a delay grows with the size of the share tree.
>>>>
>>>>Cheers,
>>>>Stephan
>>>>
>>>>Terry Lalonde wrote:
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>        
>>>>
>>>>>Interesting:  I just noticed the same thing today???
>>>>>
>>>>>02/28/2005 15:49:17|qmaster|wstlalonde|E|orders user/project
>>>>>          
>>>>>
>version
>  
>
>>>>>(88808) is not uptodate (88809) for user/project "ddodd"
>>>>>02/28/2005 15:49:32|qmaster|wstlalonde|E|orders user/project
>>>>>          
>>>>>
>version
>  
>
>>>>>(88808) is not uptodate (88809) for user/project "ddodd"
>>>>>
>>>>>repeated over and over.
>>>>>
>>>>>It's some accounting check I think.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>          
>>>>>
>>>>>>-----Original Message-----
>>>>>>From: Xavier MACHENAUD [mailto:xavier.machenaud at st.com]
>>>>>>Sent: Monday, February 28, 2005 6:34 AM
>>>>>>To: users at gridengine.sunsource.net
>>>>>>Subject: [GE users] weird "orders queue version is not uptodate"
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>            
>>>>>>
>>>>>messages
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>          
>>>>>
>>>>>>in qmaster log
>>>>>>
>>>>>>Hi,
>>>>>>
>>>>>>I'm seeing these kind of messages in my qmater log :
>>>>>>
>>>>>>02/28/2005 12:30:45|qmaster|crxu71|E|orders queue version (1668)
>>>>>>            
>>>>>>
>is
>  
>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>            
>>>>>>
>>>>>not
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>          
>>>>>
>>>>>>uptodate (1669) for queue "run at crxu76"
>>>>>>02/28/2005 12:30:51|qmaster|crxu71|E|orders queue version (1670)
>>>>>>            
>>>>>>
>is
>  
>
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>            
>>>>>>
>>>>>not
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>          
>>>>>
>>>>>>uptodate (1671) for queue "run at crxu76"
>>>>>>
>>>>>>Do you know what's the problem?
>>>>>>
>>>>>>Thanks,
>>>>>>
>>>>>>Xavier
>>>>>>
>>>>>>
>>>>>>            
>>>>>>
>>>>>--------------------------------------------------------------------
>>>>>          
>>>>>
>-
>  
>
>>>>>>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
>  
>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>          
>>>>>
>>>---------------------------------------------------------------------
>>>      
>>>
>>>>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
>>>
>>>
>>>
>>>      
>>>
>>---------------------------------------------------------------------
>>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
>
>  
>



---------------------------------------------------------------------
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