StartJob Error: Could not open plugin configuration

Deadline
1

Hi!

We’re sometimes getting a strange error.
We have a hybrid windows/linux renderfarm and this can happen on any of the two OSs.
When this happen, the affecter computer won’t ever be able to render any frame from the job.
The only workaround we’ve found is to re-submit the job.

[code]=======================================================
Error

Error in StartJob: PluginConfig: Could not open plugin configuration “/home/squeeze/Thinkbox/Deadline6/slave/render-linux-32/plugins/5413289904d26010947ec17f/MayaBatch.dlinit” because dictionary file does not exist
at Deadline.Plugins.ScriptPlugin.InitializeLocalVariables (Deadline.Jobs.Job job) [0x00000] in :0
at Deadline.Plugins.ScriptPlugin.InitializeLocalPlugin (Deadline.Jobs.Job job) [0x00000] in :0
at Deadline.Plugins.ScriptPlugin.StartJob (Deadline.Jobs.Job job, System.String& outMessage, FranticX.Processes.AbortLevel& abortLevel) [0x00000] in :0

=======================================================
Type

RenderPluginException

=======================================================
Stack Trace

at Deadline.Plugins.Plugin.StartJob (Deadline.Jobs.Job job) [0x00000] in :0
at Deadline.Slaves.SlaveRenderThread.RenderCurrentTask (Deadline.IO.TaskLogWriter tlw) [0x00000] in :0

=======================================================
Full Log

0: Task timeout is disabled.
0: Loaded job: 030 - Render.v002 (5413289904d26010947ec17f)

=======================================================
Details

Date: 09/17/2014 04:25:36
Frames: 91
Elapsed Time: 00:00:00:02
Job Submit Date: 09/12/2014 13:08:24
Job User: pouellet
Average RAM Usage: 6015170853 (37%)
Peak RAM Usage: 6015537152 (37%)
Average CPU Usage: 100%
Peak CPU Usage: 100%

=======================================================
Slave Information

Slave Name: render-linux-32
Version: v6.2.0.32 R (2563d5bc8)
Operating System: CentOS release 6.5 (Final)
Machine User: squeeze
IP Address: 192.168.1.232
MAC Address: 74:27:EA:28:1C:54
CPU Architecture: x86_64
CPUs: 8
CPU Usage: 100%
Memory Usage: 5.6 GB / 15.5 GB (35%)
Free Disk Space: 0.000 Bytes
Video Card: NVIDIA Corporation GF119 [GeForce GT 620 OEM] (rev a1)[/code]
On an affected computer, I’ve looked in the plugin folder (see log bellow) “/home/squeeze/Thinkbox/Deadline6/slave/render-linux-32/plugins/5413289904d26010947ec17f/” and it’s empty.
Deadline doesn’t seem to regenerate the content if it’s missing so that can explain why the computer won’t work for this job.
This also explain why re-submiting the job work, because the job get a new id wich force a new folder to be poppulated.

Do anybody ever encountered this problem?
Any idea why this happen and how can we prevent it?

Any help is appreciated, thanks! :slight_smile:

2

This error comes up when the plugin files, especially the dlinit file, were not copied from the repository after the slave did it’s sync. Often this can be due to network instability which showed the sync command that there was, it seems in this case, no files in the plugin directory.

3

That’s exactly what I was thinking, thanks Dwight for the explanation!
Our network is sometime unstable and we always have to double-check all network-related exceptions in our tools.

Can I enter a feature request so that deadline validate that files were copied successfully and re-try in case of errors? (or just give up with a clearer error message)
This could greatly enhance deadline stability in case with unstable networks like ours.
Also, it would save time since when it’s happening, slave just retry to job over and over without any success until their failure mechanism kick in.

Thanks!

4

Hello,

I am told such a feature, to check on the files before proceeding, is already planned for a future build, to avoid these things happening as often. Thanks for letting us know this is a priority for you!

5

Thanks a lot for the update! :slight_smile: