Hello,
Recently we have been attempting to use the Web Service REST API to collect statistics about the status of our render farm. We have a dedicated Windows machine running the Pulse 7 application. Our Windows render farm is running Deadline 7.1.2.1. We request information from the webservice every 15 minutes through a cron process. However, we are getting frequent crashes from Pulse 7 each day, and we would have to manually restart the application on the machine.
From the Deadline logs, it looks to be some timeout issue with MongoDB. I attached the error to this post.
Is there something that we can do to prevent Pulse from crashing so frequently?
Thanks in advance!
~Kacey traceback.txt (2.04 KB)
We’ve made a few fixes since the 7 days, but I’m not sure if failing connect to the DB should have ever kept it locked up.
You can see the problem specifically from the stack trace is that Pulse wasn’t able to connect to the database. I’m wondering actually if other parts of Pulse are also failing (I’m seeing that here and there lately). Could you try doing something else like house cleaning (‘control’ menu) to see if that code fails as well? That is completely separate from the web service code and should show if the problem is lower down the stack.
The House Cleaning process seems to be running fine. Though I should note that we didn’t get Pulse crashes until we began using the web REST API. Is there anything else we should test for?
That’s actually why we moved the web service out. While we spent a lot of time hardening the web server, there were cases when custom script or other oddities could bring down Pulse.
If you want, you can run a second and third Pulse on different machines and use them instead of the primary.