Managing DBR rights.

[Crazy Homeless Guy]

We are trying to find a solution in our office in regards to managing who has priority on Distributed Render in Mental Ray.

 

Basically, we have 6 users, and a bank of machines that are dedicated for DBR. Right now, our only method of deterring has priority for use of these machines is through email, stating that you have a deadline. This should be fine, but overall, we are a large company, and there are bound to be people who try and distribute to these machines when someone else is trying to hit a deadline.

 

Ideally, we would have a software solution which would allow people to check out the machines for use for DBR, based on their deadline. This would not allow access to these machines by someone who does not have them checked out.

 

I am open to all solutions short of cutting hands off of those who don't work within the verbal schedule.

[Scott Dombrowski]

Did you ever find a solution for this? I might be interested in a somewhat similar setup. In my case, I'd like to be able to harvest the idle CPUs in the building (at night, people on vacation, people gone for the morning, whatever) for both DBR for stills and network rendering through Backburner for animations. I'm thinking one way to do it would be tied into the user login, so that when a user is logged in, their PC is removed from the available pool. But, that's just a guess.

[Crazy Homeless Guy]

Nope. It is going to take someone smart to figure this one, and not a hack like myself.

[John Dollus]

those are kinda two separate questions I think. For Travis, it's a bit more complicated using DBR as the system is designed to always poll for requests from various locations so assigning priority to any one of those random requests to allow 'bumping' on a bucket level would be, well, pretty darn difficult in my mind.

As a workaround, you could create groups within backburner that get assigned to certain pools of users but, inevitably, you run into the same problem if not worse because there would be cases when cpu's from other groups would be idle but unavailable to you.

 

In Scott's case, it's a bit more cut and dried as you can create a routine which runs on a background user account and polls the CPU to see when it's idle and you can set that routine on a schedule to eliminate normal work hours. I checked into this a while back when I was in similar circumstances and this is about as far as i got with it by working with a programmer friend:

 

if you had a script that ran a WMI query it could poll the CPU to see when it is idle. If you set the script to take action when this condition is set for x time you could have it fire off your service and kill it when the CPU picks back up again. If you run a wbemtest from your run line you can start to dig into the WMI under CIMV2 namespace (under PerRawData I believe) you can start to find the naming conventions that you will need to query. A few 'if' statements and you are done.

 

I saved the text thinking it may come in handy some day

 

You could try something like http://www.rentacoder.com and find someone who can create an application for you.