AWS Thinkbox Discussion Forums

max2015 with service mode

I got the error when I use max2015 on the service mode of deadline (6.2.0.32).
“Error in StartJob: 3dsmax startup: Error getting connection from 3dsmax: Monitored managed process “3dsmaxProcess” has exited or been terminated.”

If I don’t use service mode, I can rendering correctly. And max2014 can rendering correctly even if it is on the service mode.
Only 2015+ service mode happens error.
MAX2015ERROR.txt (21.1 KB)

When you submit your job, is the scene explorer window open in Max? If so, can you try closing it and resubmitting to see if that helps?

We had tested this during the 6.2 beta to see if it was still a problem, but we couldn’t reproduce, so we updated the submitter to no longer close and re-open the scene explorer during submission. Max now docks the scene explorer, but there is no way from maxscript to close and open and re-dock it again, which prompted us to remove this from the submitter in the first place.

Cheers,
Ryan

I tried to close scene explorer, but same error still happens.

Ahh, OK, I found an answer.
If I start max on render node once time and close scene explorer, that problem will be gone.
However, it is troublesome. We have to find the switch of the scene explorer in the ini file…
Do you know that??

Found it.
We need
C:\Users\user name\AppData\Local\Autodesk\3dsMax\2015 - 64bit\ENU\en-US\UI\Workspaces\Workspace1.se.ini
And [Explorer] category and any parameter in that file. And the parameter have to set “Visible=False”

If I make that, problems not happens.
If original file have scene explorer and open(dock) it when I submit a job.

thanks

Good to know that it is related to the scene explorer. One option might be for Deadline to update every workspace.ini file to turn it off before launching 3dsmax. We’ll have to do some more thinking on this one, but glad to hear that you have a workaround for now.

Cheers,
Ryan

Hmm, We still have trouble situation.

  • If I create ini file(this is for the default Scene Explorer) and have only 1 default scene explorer, this situation is OK. We don’t need close that SE.

  • If we have 2 or more SE, like default and Layer Manager, we have to close Layer Manager and the other SE.(don’t need close default SE)
    If we don’t close Layer Manager before submit job, We get the error about “Error in Startjob: Monitored managed process “3dsmaxProcess” has exited or been terminated.” on the SERVICE MODE nodes.
    And if we close Layer Manager, we get the error about “Error in Startjob ExecuteScript: Timed out waiting for the lighting 3dsmax plugin to acknowledge the ExecuteScript command.” on the APP MODE nodes.
    But this time out error happens sometime. And if that node start rendering correctly once time, never happens on that job. But next job makes this error for this node sometime…

So I have a question. Can I control timeout value for the Lightning plugin???
And if we create ini file for the all of SE in the scene file every time when submit the job, we may avoid this problem.
But this is not good for the workstation maybe…

  • If I create the scene on the max2015 and it convert to the 2014 by save as. And I open that file in the max2014 and save it again in 2015. And next, I open this saved scene file in the max2015. And submit job from this scene file. SERVICE MODE of nodes have error about “Error in Startjob: Monitored managed process “3dsmaxProcess” has exited or been terminated.”
    In this situation, we can not avoid this error, even if I delete all of the SE. I imagine that original SE have some dirty data in the scene file, because 2014 can not read SE data from 2015. If we create new scene in the max2015 and marge all of object from that file, I can make rendering normally.
    And If I create the scene file in the 2014 and load it 2015 simply, this situation not happens the error.
    This is the autodesk problem I think.

thanks!!

“Error in Startjob ExecuteScript: Timed out waiting for the lighting 3dsmax plugin to acknowledge the ExecuteScript command.” error happens on the APP MODE node sometime too.
But this is very rare case…

Thanks for the additional info. Sounds like the problem isn’t as simple as editing the .ini files. It definitely seems like this is a 3dsmax issue, so it might be worth logging a bug with them in case we can’t find a decent workaround.

For the ExecuteScript timeout problem, you can edit 3dsmax.py in \your\repository\plugins\3dsmax to increase the timeout. The current timeout is 10 seconds. Just look for this block of code (around line 1991):

        response = ""
        try:
            response = self.MaxSocket.Receive( 10000 )
        except SimpleSocketTimeoutException:
            self.Plugin.FailRender( "ExecuteScript: Timed out waiting for the lightning 3dsmax plugin to acknowledge the ExecuteScript command.\n%s" % self.NetworkLogGet() )

Simply increment 10000 to a larger value. You could start with a minute (60000) and see if that helps. If it does, we might consider simply increasing this timeout in a future release.

Cheers,
Ryan

This is still happening with our farm and I have increased the time to 240000
Deadline Version: 6.2.0.25 R (6acea2ad2)

We run deadline as a service on our workstations and we need this working. We’ve lost several nights rendertime to this problem. Please help!

//Ryan

Autodesk has acknowledged that this is a known issue with running 3dsmax 2015 as a service:
area.autodesk.com/blogs/maxstati … -a-service

That link provides an explanation and a workaround. We’re in the process of seeing if we can automate this workaround for Deadline 7.

Cheers,
Ryan

Hi,

Please find attached an updated 3dsmax.py plugin script which attempts to automatically fix this issue.

To install:

  1. Backup your existing “…/your_deadline_repo_path/plugins/3dsmax/3dsmax.py” file
  2. Unzip and overwrite attached file with the existing 3dsmax.py file in the above path location.
  3. Restart/Requeue any failed jobs that were previously failing (no need to resubmit).

3dsmax_v6.2.1_SE_PATCH.py.zip (28 KB)

Ensure you are already running Deadline v6.2 before attempting to use this patch.

Please let us know if this fixes the issue for you / improves the situation or just plain doesn’t work? If the later, please provide error job reports.

We will include this fix in the upcoming v6.2.1 maintenance release as well once we hear back from you guys that all is well?

Regards,
Mike

Privacy | Site terms | Cookie preferences