Node and workstation not communicating properly

Render Manager Issues
#1

Hi,

I have had a consistent bug for a couple of months now.

It seems my node PC and main PC fail to communicate when one is turned off/restarted.

For example:

  • I have jobs “A, B, C” in RM with both PC’s turned on and Pulze RM running.
  • I turn off node PC
  • I remove job C and turn node PC back on.
  • Once node loads Pulze RM job C comes back automatically and is listed in both machines RM.

Similarly;

  • I set off job A on my master PC for node PC to run.
  • I shut down master PC. (let’s say when job is 50% done)
  • Node PC finishes rendering job A
  • I turn on master PC and it thinks job A is still 50% done, so it restarts the task.

Or; (this only happens sometimes)

  • I add jobs on master PC with node PC turned off.
  • I turn on node PC
  • All jobs are wiped from RM once node PC loads Pulze RM.

This is regular, mostly predictable behavior. I don’t think it used to happen like this and I’m sure it’s a bug.

I have my Pulze repository on a shared drive (hosted by master PC - so this might cause some of the issues). I have tried creating new pools with no luck.

Hopefully this can be replicated and investigated. Many thanks!

#2

I’m having a similar issue whereby if i restart pulze on my main machine it will restart the jobs rendering on the nodes as well…?

it also doesn’t accurately update the jobs… ie. a job is rendering on a node yet its listed as “waiting for nodes” in RM on both my main PC and render nodes.

I’m running the latest version 2.1.8 (2022-12-07)

#3

Hi @AnotherAngle3D and @luke

Thank you very much for the report! Could you please recreate the issue and send us some collected log from the computers that is/was rendering and the workstation after the restart, based on this guide: https://www.pulze.io/faq/how-do-i-collect-my-render-manager-logs-and-send-them-to-pulze-support and we will investigate further.

#4

Hi,

I’ve just switched my node and master PC around. So now I send off jobs from what was previously the node.

I’ll see if it still happens in this configuration or not and let you know.

#5

Still happening. Very annoying, to be honest. Please see the video, the embedded screen (bottom right) is a remote desktop cast of the node PC loading Pulze.

I have RM open on master PC, and add a job. I turn on node PC and load Pulze and it deletes the job I added (it essentially defaults back to whatever the list was when the node was turned off).

Not sure some logs would even help with this, but I’ve gathered them…

#6

@AnotherAngle3D the file and the video is no longer accessible. Could you please drop us a line to support@pulze.io and our team will have a closer look.

#7

Apologies! I somehow missed this message.

I’ll try to find the time to re-record this (I deleted it).

In the mean time, I have more information for you: as a test I changed the location of the repository. It has been on a mapped network drive (as instructed in a previous email). This was when the error was happening. Out of interest, I changed it back to a Dropbox folder (yes, I know I shouldn’t be doing this) but the error stopped happening! So it seems the reported behavior above only happens when it’s on a network (or mapped network) location.

#8

hey @luke, just wondering if you also have a mapped network drive set-up as your Pulze repository, like me?

I changed my mapped drive to a different SSD and the problem (after a quick initial test) seems to not happen any more…

Edit: Scrap that! It’s still happening as usual :frowning: