Now I could submit mission , and render them well at first, but after a while the slaves begin to Idle one after another. And after enough time, all the slaves’ status would be IDLE, even if there are so many missions to render.
I go to the slave’s output window and find the following exception are output continuosly:
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
---- August 29 2011 -- 10:18 AM ----
Scheduler Thread - Unexpected Error Occured
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
Exception Details
ArgumentOutOfRangeException -- Index and length must refer to a location within the string.
Parameter name: length
ArgumentException.ParamName: length
Exception.Data: ( )
Exception.TargetSite: System.String InternalSubStringWithChecks(Int32, Int32, Boolean)
Exception.Source: mscorlib
Exception.StackTrace:
at System.String.InternalSubStringWithChecks(Int32 startIndex, Int32 length, Boolean fAlwaysCopy)
at Deadline.Slaves.FailureTracking.FailureCache.CleanupBadJobs(StringCollection2 keepJobIds)
at Deadline.Storage.Caches.DeadlineDequeueCache.DequeueTasks(SlaveState& slaveState, String currentJobId)
at Deadline.Scheduling.SchedulerUtils.DequeueTasks(DeadlineController deadlineController, DeadlineNetworkSettings networkSettings, SlaveState& slaveState, Boolean cacheLimitGroupStubs, Boolean verbose)
at Deadline.Slaves.SlaveSchedulerThread.ThreadMain()
It seems deadline have problem when clean bad job?
Both of the C: and the RAM’s free space are enough, and there’s no problem in the file submitted.
So, I could found out what’s wrong, any help would be appreciated.
That’s an interesting error. A couple of initial questions:
Which version of Deadline are you using?
Have you recently upgraded from an older version of Deadline? If so, which version did you upgrade from?
The error seems to be a result of a job ID that is shorter than expected. We can add a check to the code to ensure this error doesn’t happen in future versions, but it would be interesting to figure out why it’s happening to you in the first place.
I worked a hard time last afternoon, I had to check the slaves and restart the idle again and again.
Hope you genius guys could find out what’s wrong with my deadline. Now I am going to create and use a new repository to test if that problem gone.
Yeah, I was wondering if there might be a “bad” file in the repo that was somehow causing this. Definitely let us know if the problem persists or not after connecting to the new repository. We have already tweaked our internal code to prevent this problem from occurring, so this issue should be obsolete in Deadline 5.1.
Thanks Ryan for your time However bad news, this problem still exists after new repository… even if the host computer of the repository has changed. I checked all the settings and couldn’t find anything strange…
I found that exception mostly come about at the time when one mission has been completed and slaves are moving on to the next to render – but not sure if conincidently.
I would keep on checking if this problem come about in some laws or not. If I solve that or find something valuable I’ll post it here.
I’m actually surprised a clean repository didn’t help. We’ve never seen this problem before, and I was thinking it was perhaps due to some corruption of the repository, but now that doesn’t appear to be the case.
One other thing to check is that all your nodes and workstations have the same version of Deadline installed. If you guys run pulse, check the pulse version too.
Hi , I found the problem has gone! (After the third repository being created)
The point cause the problem I think is, I have modified the SubmitMayaToDeadline.mel. The Maya here couldn’t pass line 663, when I open the mel the code shows like
And when execute on this line Maya think the string hasn’t been completed… So I commented it. (another"Any other have this problem"?)
When I create the 2ed new repo the error message comes about again and I just copy the first one… So I made a mistake that I thought I have created a whole new repo…
I am not sure if it is the modified mel that cause the problem, but when I created the 3rd new repo I used notepad changing the encoding of the mel instead of commenting it, and Maya recognize it well, what make me excited is the strange idle has not appeared for several days.
I’m really glad to hear that the problem is no longer occurring!
The melscript problem was a known issue in 4.1, and was fixed in version 5 (we just removed that line too). It shouldn’t be related to the ArgumentOutOfRangeException you were getting though, since it sounds like the melscript problem didn’t even allow you to submit jobs in the first place.
Hopefully things stay stable for you now, but of course if any issues creep up, just let us know!
However bad news, this problem still exists after new repository… even if the host computer of the repository has changed. I checked all the settings and couldn’t find anything strange…