Vray 1.5 finale : Backburner and DR

[raistlin]

Raistlin, the list of DR servers is not submitted along with your scene; it is local to the machine which initiates the DR rendering. So, no, you will not be able to use this in the way that you want. We will consider it for future builds though.

Best regards,
Vlado

ok i want to tell you that it's really usefull to render with the DR and backburner, Thanks to the team to all these good news !

[panthon]

It's exact location varies with your particular system, so you may have to search for it.

At last ive managed to find that file! For some reason windows flag some user folders as "hidden", so a regular search wasnt able to find the DR file, which happened to be inside one of these hidden folders. You must tell the search engine to look for hidden files or hidden folders too (this option is unchecked by default) by going to advanced parameters and checking the option.
Telling windows folder to show hidden files is NOT enough since a search wont find them. I hope its useful for some other users!

Now DR through Backburner works!

[beestee]

If I understand correctly, the settings in the "vray_dr.cfg" file will be used any time a DR job is sent to that node through backburner, regardless of the on/off states that I sent the file with. This is acceptable because at any point I would be using the full processor potential of my farm.

But I do have a few items that could be added to the wishlist.

1. It would be nice if the DR spawners were aware of backburner server status, so that I wouldn't be able to inadvertently put my nodes under double duty.

2. I have not tested this, but are the DR spawners aware of the backburner priority setting? If not, it would be a nice feature so that they could abandon a current low-priority job and let the host continue it, then pick up on a job that was set to a more urgent priority until it is finished, then look for the next job in order of priority.

[vlado]

For the 1st point, a slave that already works on a DR job, will not accept another job until it finishes the first one.

For the 2nd, it would certainly be a useful addition and we will consider it. Thanks for the suggestion!

Best regards,
Vlado

[raistlin]

Sometimes we have few nodes render which connect faster than the host and systematically don't render proxy or don't render any maps... Is there a buckets limit for the DR ? We launch 3 jobs through the backburner and we have to launch a lot them to have a render without black square...

If you know where is the solution to this, i'm very interested

[vlado]

You can turn on the "Check for missing files" option in the System rollout of V-Ray; in that case a slave will refuse to render if there is any problem with missing maps or proxies, so at least you will not get black buckets.

Best regards,
Vlado

[Da_elf]

well. DR with BB is working like a dream now

[raistlin]

ok ! i'll try this new option today, i notice a really good news : the connection of a Dr spawner after the render launch work perfectly !

I m agree with beestee and a "monitor" like backburner should be very usefull, we could maybe drive the farm dr spawner before it, it must be a very very interseting tool for the next release.

[panthon]

1. It would be nice if the DR spawners were aware of backburner server status, so that I wouldn't be able to inadvertently put my nodes under double duty.

Yes, I second that. Currently, a machine can render 2 tasks at the same time:

- Contributing to a DR when its called by the client machine.

- Rendering its own task assigned from backburner

That is certainly not desirable. It´s nice that a machine cannot render 2 DR tasks at the same time, but a stronger connection between Backburner and DR would be increasingly useful.

[themaxxer]

DR over BB works fine here.

I have another question. Is it true that the server who gets the job must have a full license of vray? I read that somewhere in the forum.

another question...
do I have to modify the vray_dr.cfg every time on the server (who gets the job over BB) or can I put a 1 for all render computers and save this for all new jobs? I mean only the vray_dr.cfg on my workstation would be changed after I sent the job but not the file on the server, wouldn't it?

best regards
themaxxer

[Clifton Santiago]

I have another question. Is it true that the server who gets the job must have a full license of vray? I read that somewhere in the forum

No, thats not true. I always send my DR via BB to one particular "host" render node and it just has the render node install.

do I have to modify the vray_dr.cfg every time on the server (who gets the job over BB) or can I put a 1 for all render computers and save this for all new jobs? I mean only the vray_dr.cfg on my workstation would be changed after I sent the job but not the file on the server, wouldn't it?

I just set up the vray_dr.cfg on my host render node to include all available nodes and workstations. The vray spawners on our worstations are off by default, so if I want them included, I run a script that enables their spawner so they will participate in the DR.

Also, it does not matter which machines you specify on your local workstation before you send it, as long as DR is enabled. It will defer to the host node cfg file.

[muzzy]

For remote machines any easier way to run spawner without any remote connection to eachone one by one ?

[Clifton Santiago]

Yes, you run a script that starts or stops the service on all nodes.

We use this utility:

http://technet.microsoft.com/en-us/s.../bb897542.aspx

And then run a .bat file that looks like this, just substituting the proper node names and admin username and passord:

echo off

C:\SwitchDR\psservice \\NODE01 -u username -p password start "VRaySpawner 90"
C:\SwitchDR\psservice \\NODE02 -u username -p password start "VRaySpawner 90"
C:\SwitchDR\psservice \\NODE03 -u username -p password start "VRaySpawner 90"

...etc. to include all nodes

[3LP]

well i haven't understand all the manipulation here.

To launch a DR with backburner, i have to launch DR Spowner on all the nodes and lunch the server.exe as well ?

So i submit my bb-job to a specified client that will (thrue his own DR list) lunch the frame with all the DR that are connected. Is that true ?

Or have we only to lunch bb server and when a job is submitted with DR enable, the bb server on all the node will lunch automaticly DR spawner?

i don't understand the imbrication of all the process, could somebody help me? i want to understand why everybody is putting there hand in the air and screaming

[Clifton Santiago]

To launch a DR with backburner, i have to launch DR Spowner on all the nodes and lunch the server.exe as well ?

It depends on how you want to do it. In our case, the service VraySpawner 90 is running on all nodes at all times.

So i submit my bb-job to a specified client that will (thrue his own DR list) lunch the frame with all the DR that are connected. Is that true ?

If I understand your question, yes. We have just set up a single host node, that is set to send the DR to all other nodes.

i don't understand the imbrication of all the process, could somebody help me? i want to understand why everybody is putting there hand in the air and screaming

I don't understand this question. Are you asking for help, or complaining that too many people are trying to help? There are many different ways to set up network rendering so there are going to be many opinions on the best way to do it.