From: "Jon Herman" <jonh@(email surpressed)>
Subject: Logging in Rush and XSI
   Date: Thu, 22 Dec 2005 12:49:23 -0800
Msg# 1155
View Complete Thread (2 articles) | All Threads
Last Next
Hi,

What would be the best place to store rush log files on a render farm with a 
master submit host?
Currently, we have a logging server and a separate submits host server.
Would it be more efficient to have the logs stored on the submit host versus 
having them stored on another server?

I'm also looking for documentation for XSI logging levels and logging 
output.
Specifically, I need to know the details of  XSI progressive logging.
Below is snippet of an XSI log at the "progressive" level:
JOB  0.3  progr:    99.4%    rendered on Render1.3
JOB  0.4  progr:    99.7%    rendered on Render1.4
JOB  0.3  progr:   100.0%    rendered on Render1.3
What does it mean when it says rendered on Render1.3 versus rendered on 
Render1.4?
If there's a place I could find documentation on logging levels, please 
direct me. 



   From: Greg Ercolano <erco@(email surpressed)>
Subject: Re: Logging in Rush and XSI
   Date: Thu, 22 Dec 2005 13:40:10 -0800
Msg# 1156
View Complete Thread (2 articles) | All Threads
Last Next
Jon Herman wrote:
> What would be the best place to store rush log files on a render farm with a 
> master submit host?

	'Somewhere on your file server' is my initial response; the pathname
	would be up to you.

	I might need more info to know what you're really looking for.
	
> Currently, we have a logging server and a separate submits host server.
> Would it be more efficient to have the logs stored on the submit host versus 
> having them stored on another server?

	The config you have sounds efficient -- you've dedicated a server
	to the managing of log files separate from the single submit host server.

	As always, I warn everyone with a large network to NOT submit all
	jobs to a single server; you totally loose the distributed load management,
	and put a single point of failure on all your jobs.

	The default is for the user's workstation to be the submit host,
	and if you want to avoid that, then set up a few servers to be
	used in rotation for job serving, depending on the scale of your
	network.

> I'm also looking for documentation for XSI logging levels and logging 
> output. Specifically, I need to know the details of  XSI progressive logging.
> Below is snippet of an XSI log at the "progressive" level:
> JOB  0.3  progr:    99.4%    rendered on Render1.3
> JOB  0.4  progr:    99.7%    rendered on Render1.4
> JOB  0.3  progr:   100.0%    rendered on Render1.3

	This looks not like XSI, but Mental Ray logging.

	I don't know offhand how XSI tells mental ray what the logging level
	should be. It's possible it's controlled in the XSI scene file.

	XSI does have a command line argument "-verbose", which according
	to the minimal docs I have on XSI:

		-verbose [ "none" | "error" | "progression" | "full" | "all" ]

	..so try '-verbose error' and see if that helps.
	It looks like you might have 'full' or 'all' currently.

	But as I say, the above output excerpt you're showing appears
	to be the output of Mental Ray, so unless XSI passes the -verbose
	setting on to mental ray, you may have to dig deeper.

	If nothing else, you can filter away these 'progr:' messages by
	piping the output of XSI through 'grep -v "  progr: "' (unix)
	or 'findstr /V progr:' (windows)

	Such hacks can be made in your render script.