-
Compressor/Qmaster/Cluster failing
Hi everyone,
First of all this post is in regards to using clusters, as set up through Qmaster, for encoding with Compressor 3.0.
Okay, so basically this is the story, I’m a Film student who has been working away at the tech staff trying to get them to set up a cluster for encoding/rending material with compressor at film school.
eventually i was successful at this, and no doubt it was always going to happen anyway, however both myself and the tech staff have been running tests with the cluster (/s at this time various clusters have been set up, removed/replaced).
Most of the bugs have been worked out of the system and the current cluster works, however one thing that neither myself or the tech staff have been able to find a solution to is, while the cluster works we are getting about a 70% success rate, the 30% fail rate seems to occur when the cluster encounters a computer that is actively running Final Cut Pro.
I have personally confirmed this a number of times by seeing which computer has failed – through Batch Monitor – then gone and found another student utilising Final Cut Pro (or Motion) on the computer in question.So, does anyone have any ideas as to how to fix this? why it might be happening? Etc
-
1 Reply
-
Couple of things: Distributed rendering “reaches out” and looks for systems with available CPU cycles to speed along the process. If someone is actively using FCP, and you’ve designated that computer as “available” then you’re stretching the resources of that CPU and it may cause a failure.
One way to avoid this is to NOT designate CPU’s as part of a cluster if you know that they will be in heavy use. When we set up our cluster, we specifically EXCLUDE certain computers that we know are going to be in heavy use (The GRFX or On-line systems, for example).
Here’s a tip. “Apple Remote Desktop” is extremely helpful in “resetting” the cluster.
Mark
Reply to this Discussion! Login or Sign Up