Announcement

Collapse
No announcement yet.

network, permissions, etc.

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • network, permissions, etc.

    Okay, sorry about posting another DR thread, but...

    I finally have DR working quite well, with one hitch.
    I have to put all the resources/maps/proxies on the slave's local disk or it will not see them.

    some facts:
    -I only have 2 machines (XP Pro 64 and Vista Ultimate 64)

    -Both are administrator

    -I can freely copy, move ,delete, edit any files in either direction between the 2 machines.

    -if any resources are pathed to the non-slave, the slave renders black or empty buckets.

    -permissions on shared drives are set to everyone=full control

    DR is great, but this is a big pain in the work flow since sometimes I'm copying LOTS of files over to the slave from a shared network HD attached to the main computer.

    What am I missing? Is this a Winblows issue?

    Thanks,
    JR

  • #2
    I've been keeping all maps locally on each machine set to the following directory:

    c:\maps

    Permissions are fully available on every comp so copying between comps is not a problem. I've made all the maps directories shareable. Also, make sure Max path directories (within the program) are configured to see c:\maps.

    I have a directory on my desktop called: \slaves\. In it is a couple of batch files which makes copying my maps very easy. I copy all the maps into the \slaves\ directory then click the batch file. The batch file is called copy.bat. I've posted the code below:

    Code:
    copy *.* \\slave1\maps /Y
    copy *.* \\slave2\maps /Y
    and so on.

    Of course, change the comp name to whatever name you are using.
    I also keep the checkbox for restarting slaves upon render end checked - that seems to help a lot.

    I know this seems like an extra step, but when you have several comps all pulling large amounts of data from a single source, bottlenecks occur and things go wrong. For me this has been the most reliable method so far.
    LunarStudio Architectural Renderings
    HDRSource HDR & sIBL Libraries
    Lunarlog - LunarStudio and HDRSource Blog

    Comment


    • #3
      That sounds logical and I'll try it, but my problem is that most of my maps are on a network drive named "K" so, must I copy all the maps to local "C" drives on both computers, every time I add one to the working scene? I feel luike that's pretty much what I'm trying to avoid. Again, all these directories are shared with full permissions, etc. the main reason is the K drive is 500GB, while the C drives are each 74GB. If i keep copying all over the place, I'm soon going to have disk space problems.

      Why can't the slave use maps in my shared directory even though it has full access and permissions? Slave can move, copy, create and delete files from the shared "K" drive, but, not with DR/Max...

      This seems really dumb, but maybe it's me and my non-existent IT team
      Thanks for that tip though!
      The only hint I can ever get, is by turning on "check for missing maps", and it just says aborted, missing maps found...not much help for solving the root of the problem...

      Comment


      • #4
        Hey there; I know exactly how you fell. Being a small company with limited resources can be stressful so here is what I have found works for us.

        We have a main machine/workhorse that acts as our server and a small renderfarm composed of 4 Boxx rendernodes. We have our drives setup like this on the main server:

        * Drive letters are just examples
        F:/Projects
        G:/Resources

        We then map both of these drives on any rendernodes/additional computers on the network. As long as you give these mapped drives the exact same drive letters (F,G respectively), things will work. So every machine can read/write to these directories which are on the main server. For simplicities sake, we setup the permissions on each rendernode/workstation using the same account and set it to reconnect at logon using the same username/password. That way it'll logon automatically and you don't have to set the password every time your machines bootup.

        That should pretty much do it. As long as each of your machines can read/write to your project directories and access all plugins/textures needed by your scene it should work without any hitches.

        For plugins it can be a little more in-depth... you may have to alter your plugin.ini file to point to a network folder where you keep plugins etc. This can be a little work but once setup it really helps with streamlining to render process as now you only have to update one folder with any new plugins/scripts and all external machines will be able to access them.

        If you are unclear about this just let me know as I may have missed a step.


        cheers
        ___________________
        joel lelievre
        www.intrinsia.net
        www.delicatemachines.com

        Comment


        • #5
          THANK YOU Intrinsia!

          This tip got me to the root of my problem.

          "For simplicities sake, we setup the permissions on each rendernode/workstation using the same account and set it to reconnect at logon using the same username/password."

          I went to check how mine was mapped, and it was set to logon as the slave. I changed it to logon as the non-slave, and now it's seeing all my maps on the shared network drive!

          Thank you x10
          It was a small detail, but I never would have checked that part if not for your post.

          Cheers

          Comment

          Working...
          X