Pulse soft crash

Deadline
1

Haven’t figured out what is going on quite yet, but about once every week or so Pulse seems to just kinda stop working correctly. It doesn’t fully crash, but people who load Monitor up wind up getting incorrect info all over the place, jobs never leave the queue, they continue to show up as rendering, weirdness all around basically.

It happens at all of our sites running Pulse (with their own repositories) too. Site 1: 12 nodes, 5 workstations, crashes every 1-2 weeks. Site 2: 70 rendernodes, ~20 workstations, crashes every 1-2 weeks here as well…

We are running: Deadline 5.2.0.47700 and Max 2012 / Nuke 6.3v7 / Fusion 6.2 across the farms.

I have noticed in the past that corrupted jobs stuck in the queue tend to cause problems, would this be related? Is there a way to just have the repo nuke those corrupted jobs automatically or search them out and notify an admin or something?

J

2

Hey Joe,

Which OS do you have Pulse running on in both locations? If Pulse is running on a non-server edition of Windows, that can cause unexpected results because of Windows’ connection limitation.

Cheers,

  • Ryan
3

Windows Server 2008 for the Pulse server, Windows Server 2003 for the Repo (not sure if that would cause problems).

J

4

Nope, that shouldn’t cause any problems.

The next time this happens, can you send us the most recent Pulse log from the current session? You can find the logs from the Pulse UI by selecting Help -> Explore Log Folder. We can take a look to see if it ran into problems before things started acting up.

I should note that this won’t be a problem in Deadline 6. Pulse will no longer act as a proxy server, so caching issues like this won’t come up.

Cheers,

  • Ryan