Hi guys, I'm a pretty new Houdini user (coming from Max and Maya) and we are introducing Houdini into our pipeline. I was just curious to know the best way of setting this up as far as how much we can leave on the servers, what needs to stay on the local machines and whats the best way of managing site wide tools/otls/settings etc. I have seen some talk of the $HSITE variable but as im a bit of a noob im not sure to what extent it can be used, or if there are better ways of going about this.
We are a mid sized studio working on Windows and have a decent set of general pipeline tools made in pyton and use Deadline for our render farm.

Thanks in advance for any help!

Houdini will run just fine from a network disk. However, with windows you'll need to install the vcredistributable packages on those machines (which is part of the windows installer), which kind of defeats the purpose of having a network copy - so you don't have to install to all the machines.

However once you've done this, you can update to a dot-release and just run Houdini from the network location and avoid a mass install.

Be warned, we found that in previous versions, the permissions weren't locked down on the installs, and allowed non-admin users to inadvertently modify the contents of the hfs directory - which is pretty a much a no no IMHO. They have supposedly fixed this in the new 12.x releases, though I haven't personally verified this.

It would be nice if you could just shove the vcredist stuff in the houdini directory so you could just roll out one copy, but I think that would require some work on sidefx's part - RFE maybe…

Hi,
we also use the $HSITE variable to access our tools and digital assets. Here are two threads that helped me in setting it up correctly:

http://forums.odforce.net/index.php?/topic/11737-configuring-hip-and-a-project-directory/page__hl__hsite [forums.odforce.net]

http://www.sidefx.com/index.php?option=com_forum&Itemid=172&page=viewtopic&t=17837&highlight=hsite [sidefx.com]

Currently I'm also trying to simplify the way our software gets installed/run on our machines. I set up a startup shell script which initializes all the necessary environment variables and runs the correct software version from a remote server. This works fine on linux (the load times are longer, but not too bad) but on a mac it seems to take forever until a software starts (e.g. Nuke) and it generates a tremendous amount of network traffic. Also I could not yet get Houdini to start on a mac from a network drive.

Can someone with more experience share some tricks and knowledge how to set something like this up? Or is there a better way to do it?

Best,
-dennis