I remember there being an option not to copy the main scene file to the repository / jobs folder. Based on the built in maya submission script, i have found that if i define SceneFile in the job file, and then omit the scene name in the command line, that will make it skip the copy to the repository.
This does work, the file is now not copied. However, the job now errors somewhere in deadline:
Error in RenderTasks: TypeError : cannot concatenate ‘str’ and ‘NoneType’ objects (Python.Runtime.PythonException)
[’ File “none”, line 168, in RenderTasks\n’]
at Deadline.Plugins.ScriptPlugin.RenderTasks(String taskId, Int32 startFrame, Int32 endFrame, String& outMessage, AbortLevel& abortLevel)
Any ideas? The file loads proper, but then the failure happens here:
0: STDOUT: File read in 2 seconds.
0: STDOUT: File read in 2 seconds.
0: STDOUT: Result: //inferno2/projects/common/pipeline/submission_queue/2013_01_29/TST_003_0002_maya_animation_layout.ma_16_49_34_207/TST_003_0002_maya_animation_layout.ma
0: STDOUT: mel: READY FOR INPUT
0: INFO: This is a script job
0: INFO: Reading Plugin Info
0: INFO: Building globals melscript
0: INFO: Building DeadlineValue
=======================================================
Type
The crash appears to happen in mayaBatch on this line:
scriptBuilder.AppendLine(" case “DataFileName” : return “” + self.GetDataFilename() + “”; break;")
Auxiliary File #1 ("\inferno2\projects\common\pipeline\submission_queue\2013_01_30\TST_003_0002_maya_animation_la
yout.ma_09_42_14_521\MayaScriptJob_0_job.job")
Error: An error occurred while deserializing the Pools property of class Deadline.Pools.PoolCollections: Enumerable clas
s PoolCollection does not implement IList so it can’t be deserialized. (System.IO.FileFormatException)
at MongoDB.Bson.Serialization.BsonClassMapSerializer.DeserializeMember(BsonReader bsonReader, Object obj, BsonMemberM
ap memberMap)
at MongoDB.Bson.Serialization.BsonClassMapSerializer.Deserialize(BsonReader bsonReader, Type nominalType, Type actual
Type, IBsonSerializationOptions options)
at MongoDB.Bson.Serialization.BsonClassMapSerializer.Deserialize(BsonReader bsonReader, Type nominalType, IBsonSerial
izationOptions options)
at MongoDB.Driver.Internal.MongoReplyMessage1.ReadFrom(BsonBuffer buffer, IBsonSerializationOptions serializationOpt ions) at MongoDB.Driver.Internal.MongoConnection.ReceiveMessage[TDocument](BsonBinaryReaderSettings readerSettings, IBsonSe rializationOptions serializationOptions) at MongoDB.Driver.MongoCursorEnumerator1.GetReply(MongoConnection connection, MongoRequestMessage message)
at MongoDB.Driver.MongoCursorEnumerator1.GetFirst() at MongoDB.Driver.MongoCursorEnumerator1.MoveNext()
at System.Linq.Enumerable.FirstOrDefault[TSource](IEnumerable1 source) at Deadline.StorageDB.MongoDB.MongoSettingsStorage.GetPoolsAndGroupCollections(Boolean invalidateCache, PoolCollectio ns& poolCollections, PoolCollections& groupCollections) at Deadline.StorageDB.SettingsStorage.GetActivePoolsAndGroups(Boolean invalidateCache, PoolCollection& activePools, P oolCollection& activeGroups) at Deadline.Controllers.DataController.SubmitJob(Job job, StringCollection2 auxiliarySubmissionFileNames, Boolean loa dPluginInfoDictionary, Boolean resetJobDefaults, TaskCollection taskCollectionOverride) at Deadline.Submission.SubmissionUtils.SubmitNewJob(IDictionary2 htInfo, StringCollection2 auxillarySubmissionFileNa
mes)
at Deadline.Submission.SubmissionUtils.SubmitNewJob(String[] args)
at Deadline.Submission.Submit.Perform(String[] args)
Before this, i assigned a new machine to a pool / group.
But now it appears the mongo db is corrupted. Thats pretty scary.
I dont see any pools listed in the monitor’s edit pools dropdown (neither are any slaves listed). If i go to “edit groups”, it pops up the Manage Pools window, still empty.
The manage pools dialog is completely inresponsive, none of the buttons do anything other than the little x in the corner to close the window.
I dont know how to fix the issue…
Changing pools/groups is a pretty simple and common operation, i really hope its not this easy to corrupt the database (we are currently unable to submit any new jobs)
Thanks Ryan, thats great news. We got most of the code changes done last night for our intended rollout this morning, so you might imagine this being a bit of an upset :-\
I read on the bug forum that this seems to be a mongo driver issue? While you are doing the build, i might try swapping out dlls, if that is something that might work. If not, we will wait.
Btw, the manage pools dialog was fine when i first used it with beta 10. Same for managing groups. So whatever was set up with beta8 was still visible this morning.
Then i edited the pool settings on one machine, and clicked ok. After that going to edit groups still showed the valid group settings. But after editing the group on the machine, both edit pool/edit group dialogs came up empty.
Anyway, maybe this helps, unless you already know whats happening.
Yeah, the issue was that beta 10 could read beta 8’s pools, but it can’t read beta 10’s pools. So basically, once beta 10 “updated” the pools, it could no longer read them back in.
It’s already fixed internally, and the beta 11 build is almost done.
Please read the release notes, as there is still a bug affecting the saving of the pools and groups. It is not a critical bug like the one in beta 10 though, and it will be fixed in beta 12.
(