Re: DrQueue

["Alan Jones" <skyphyr(at)gmail.com>]

Hi Stefan,

It's designed so it'll be (relatively) easy to support other programs too.

You can also add custom environment variables per job if you require.

Oh and while I've not tested on windows, the code I've written should
be cross platform already. I've put a but of code designed to make the
things which might not be cross platform work cross platform. Plus the
database is designed to handle OS specific jobs already - though not
to distribute any given job among multiple OSes (handling path issues
didn't seem worthwhile - especially given a farm is generally one os
for a particular job.

Aiming to have the first version ready around 10th August.

Cheers,

Alan.

On 7/27/06, Stefan Andersson <sanders3d(at)gmail.com> wrote:
> Great! I will look forward to see where this leads. Would "Splish" support
> other kinds of network things too?
>
> regards
> stefan
>
>
>
> On 7/27/06, Alan Jones <skyphyr(at)gmail.com> wrote:
> >
>  Hi Stefan,
>
> We're currently running DrQueue here. I've customized it to support
> XSI however I'm currently writing a new renderfarm from scratch here.
> I've just finished the backend and it's in testing. Starting on the
> submission programs and then the gui.
>
> We're open sourcing it - the project page is
> http://sourceforge.net/projects/splish
>
> Splish can manage jobs by priority and user (say all users had
> multiple jobs on equal priority - the farm would be split evenly among
> them - so 4 users would get a quater of the farm each - then once one
> of them was finished a third each.) Higher priority would naturally
>  take the percent it requested. So you may have priority 100 for 30% of
> the farm (number of machines is expressed as a percentage of the
> available machines).
>
> Fail handling is fairly advanced in it too. If you submitted in
> batches of 10 and the render dies after 4 frames - it will realize and
> create a new task to render the remaining 6 frames - making the four
> it completed as done. If the first frame fails then it is only
> attempted twice more before marked as failed. Any machine which fails
> on a job will no longer attempt that job.
>
> It can stream logs in realtime to any machine you like - i.e. as fast
> as you get it when rendering locally.
>
> If you've got any questions feel free to grab me on msn or something.
> skyphyr_(at)hotmail.com
>
> Cheers,
>
> Alan.
>
> On 7/27/06, Stefan Andersson <sanders3d(at)gmail.com> wrote:
> > Hello,
> > I was wondering if anyone has tested DrQueue with XSI?
> >
> > http://www.drqueue.org/
> >
> > Looks a bit messy but has some promising features.
> >
> > regards
> > stefan
> >
> >
> > --
> > - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
> > Freelance Pixel Abuser
> > http://os3d.blogspot.com/
> ---
> Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
> unsubscribe xsi
>
>
>
> --
> - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
>
> Freelance Pixel Abuser
> http://os3d.blogspot.com/
---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
unsubscribe xsi