Hi when i render in Maya batch render with “Arnold stand in” with animation which works with file sequence and the file extensions inside maya are read as followed- “test.####.ass”. For some reason it cannot render, but if I don’t use “stand in” with sequence(just one frame ass for stand-in) then it renders just fine . What could be the solution of network rendering Arnold stand-in sequence animation with deadline .
Please note that the whole project folder where the stand-in sequence files are stored are shared in the network. As you can see from the image the stand-in ass sequence cannot be copied from network to
C:/Users/…AppData/…/ jobData/…
folder for some reason. But if there is no sequence , rather just one stand-in ass file then it can copy in the above folder and renders fine. But that defeats the purpose of Arnold stand-ins with animations. Question is why can’t it render sequence? please help!
Also note that the pc from where the project folder is shared, that pc slave can render the sequence but other pc cannot giving the error as given on the pic.
I can provide any information to figure the problem out. No resource at all on the internet regarding using arnold stand-in (proxy) sequence animation to run with deadline. You guys are the only hope!
Hmm. Are you choosing the “Submit Scene File” option in the submitter? It looks like the stand-ins may be relative to the scene file, and the scene file option will copy the scene to the Repository, then to the render node which will break any and all relative paths.
Luckily, since Deadline 7.2 we’ve supported remapping the scene without making a copy of it so cross-platform rendering won’t be a problem. Nuke, AE, and V-Ray standalone all create a local copy for path mapping.
In short, if you’re using “Submit Scene File”, turn it off and see if it helps.
i didnt untick submit scene file…as its 9pm right now in here i will try it next day and post here what happens…the full job reports… … please stick with us on this, our studio is hanging on this problem… as we are totally dependent on deadline for our renders…also for note… rendering in arnold export job shows the same error…
when i render the scene with single frame of stand in file everyone on the network can render, but when i use sequence of stand-in in Maya and render, only the server pc slave can render and other pc shows error(even if i untick submit scene file). Here is my job report of that:
i managed to find a solution somehow but fell in few other problems. The solution is setting the Arnold stand-in path absolute with network path like “”//walid/morpheus&malcolm/speedTree/tree_proxy/palm_tree01.####.ass" replacing “D:/work//morpheus&malcolm/speedTree/tree_proxy/palm_tree01.####.ass”, this way it renders in Maya batch render via network PC, but takes a long long time to render the first frame, because from the log i can see it was copying 960 .ass files( my animation is 960 frames). Secondly, if i render the same file in Arnold export job, it successfully generate the ass file taking long time(because of copying), but when rendering the image it shows error "The given path’s format is not supported. " I attached both the reports here. Now the problems and questions i am facing are:
When i had stand-in path like "D:/work//morpheus&malcolm/speedTree/tree_proxy/palm_tree01.####.ass, without sequence it was working(copying the single .ass file to cache and rendering fine), but this does not work using sequence. Why?
The copying process of all the stand-in ass file will be just for the first frame? Will other frames render fine?
Why can’t i render using Arnold export job?
Also if there any other way of avoiding the initial path remapping by copying? It takes a long time to copy these 960 .ass files( folder size-17 GB for just one stand-in) to the cache directory. For some reason if any task is re-queued the whole 960 .ass copying process starts again before rendering that task!
Okay, so I have a fair amount of background on this now. Sorry it took some time to really wrap my head around the system.
So, the mechanism as it exists now doesn’t have a good way to compare single AOVs with a group of them. Also, we have no idea how many tasks a Slave will pick up during its run and so we must remap all the AOVs up front. In chatting with the dev team, it may be possible to change the mappings on-demand but it will take some developer effort.
If you do not have to change the paths for the AOVs and they will always be in the same folder, you can backup the file at “[repo]/plugins/MayaBatch/MayaBatch.py” then find this line:
if self.Renderer in [ "arnold", "arnoldexport"]:
scriptBuilder.AppendLine( 'catch( python( "DeadlineMayaBatchFunctions.arnoldPathmapStandin(\'' + self.TempThreadDirectory.replace("\\","/") + '\' )" ) );' )
and change the first line to read:
if self.Renderer in [ "arnold", "arnoldexport"] and False:
To disable the AOV mapping until we find a more elegant solution.
