I'm having problems with Houdini and both local and remote rendering. When I attempt to render locally on a Quad processor system using all 4 cores Houdini will crash instantly with these errors:


Successfully loaded: /home/doug/project/milbank/hip/ves0.hip


Rendering Object: /obj/RBC Frames: 1 -> 400 At: 04/27/07_16:35:02
To: /project/milbank/pic/ves/rbc/tga File: rbc.tga On: drip,drip,drip
mantra: Network Error Unable to contact remote hserver
mantra: Network Error Unable to contact remote hserver
mantra: Network Error Unable to contact remote hserver

mantra: No remote hosts are available for rendering
Render failed:
Error: Could not write out to the script or pipe
Failed to complete render: mantra1
done At: 04/27/07_16:36:02

real 1m5.242s
user 0m9.921s
sys 0m11.353s


As you can see it only takes about 10 seconds for it to error out at the console. Also while using less cores it takes longer to error out.

While remote rendering on multiple dual core computers it takes around half a hour or so for one of them to crash. Normally when doing this on the same 2 remote rendering computers the same one always errors out and hserver will crash on it. The workstation will come up with a error but the system that hserver crashes on i cant seem to find anything that says why it would.

rendering Object: /obj/CapSurfaceSick Frames: 118 -> 260 At: 04/27/07_11:52:49
To: /project/milbank/pic/vesfast/capsurfsic/tga File: cs.tga On: fog,fog,dew,dew
mantra: Network Error Missing data in read – possible remote crash
UT_NetPacket::read: Connection reset by peer
mantra: Network Error Missing data in read – possible remote crash
UT_NetPacket::read: Connection reset by peer
mantra: Network Error Missing data in read – possible remote crash
mantra: Network Error Missing data in read – possible remote crash



Thanks for any help you can provide, its appreciated.

Hi Indarios

We have setup a renderfarm just recently and we had the same problems.
Fix the host file in /etc/hosts as such:

127.0.0.1 localhost
(your local ip adress) drip
(rendernode ipadress) name

It is important that you don't mix or alter the localhost line.
You have to have the same file on all machines so they can all contact each other.
You also have to have setup Hserver in the hserver.opt file, there is more to read about this under
http://www.sidefx.com/index.php?option=com_content&task=view&id=702&Itemid=229 [sidefx.com]
Hserver has to be running on all rendernodes and you have.
If you are running machines on diffrent subnetworks the mask have to be changed +.+.+.*, but this is over my head.

Hope this help

Hello jmpr@mac.com

Thanks for your reply, but what you suggested didn't seem to have any affect as of yet. I double checked my hosts files and they all seem to be ok. I am also messing around with the hserver.opt file but everything I do to it seems to yield no results.

Currently my workstation (drip) does render on itself and remotely but I figured out that depending on cpu power and time that hserver will crash and i will start getting the "mantra: Network Error Unable to contact remote hserver". Then ill start hserver again and try rendering but the same thing comes up(this is when I render just on drip. I noticed that this happens on one of our nodes as well, but I'm not sure about the others.

Since I have been messing with this I seemed to have been able to get my remote machines to render from my workstation without hserver crashing by simply removing most alias from the user account. When I tried this from my workstation It now seems to render just a tad bit longer before hserver crashes.

While rendering on my workstation and on all 3 nodes hserver on my workstation does still crash but comes back up and renders for about half a minute before it crashes again for a minute or so, and while this happens all of my nodes will stop rendering as well. Very strange indeed… anyone have any ideas to what my problem might be?