cappie
09-13-2004, 08:20 AM
I have a couple of questions:
1) Is it possible to start LWSN.EXE as a system service at low priority?
I'd like to build a renderfarm using idle cycles on workstations, instead of having dedicated hardware for it... and it would be nice to have it a bit crash-insensitive
2) How to I say in a 'job'-file that a certain frame has to be re-done after a crash or something? (to check for missing frames)
3) Will Screamernet III support IP based communication or something else non-directory-based?
For anyone who is interrested to make their screamernet somewhat failsafe (except for missing frames after crashes):
l: = L: (drive mapping to shared Lightwave directory)
@echo off
echo --= STARTUP NODE %1 =-- -- - > l:\sn_commands\node%1.log
echo LWSN NODE %1 started at: >> l:\sn_commands\node%1.log
date /T >> l:\sn_commands\node%1.log
time /T >> l:\sn_commands\node%1.log
echo ---- --- -- - -- - >> l:\sn_commands\node%1.log
echo init > l:\sn_commands\job%1
goto start
:crashed
echo --= CRASHED NODE %1 =-- -- - >> l:\sn_commands\node%1.log
echo LWSN NODE %1 crashed at: >> l:\sn_commands\node%1.log
date /T >> l:\sn_commands\node%1.log
time /T >> l:\sn_commands\node%1.log
echo ---- --- -- - -- - >> l:\sn_commands\node%1.log
goto start
:start
@LWSN.exe -2 -dl:\ -cl:\sn_config l:\sn_commands\job%1 l:\sn_commands\ack%1
goto exit
:exit
if NOT %ERRORLEV% == 0 GOTO crashed
echo --= STOPPED NODE %1 =-- -- - >> l:\sn_commands\node%1.log
echo LWSN NODE %1 stopped at: >> l:\sn_commands\node%1.log
date /T >> l:\sn_commands\node%1.log
time /T >> l:\sn_commands\node%1.log
echo ---- --- -- - -- - >> l:\sn_commands\node%1.log
1) Is it possible to start LWSN.EXE as a system service at low priority?
I'd like to build a renderfarm using idle cycles on workstations, instead of having dedicated hardware for it... and it would be nice to have it a bit crash-insensitive
2) How to I say in a 'job'-file that a certain frame has to be re-done after a crash or something? (to check for missing frames)
3) Will Screamernet III support IP based communication or something else non-directory-based?
For anyone who is interrested to make their screamernet somewhat failsafe (except for missing frames after crashes):
l: = L: (drive mapping to shared Lightwave directory)
@echo off
echo --= STARTUP NODE %1 =-- -- - > l:\sn_commands\node%1.log
echo LWSN NODE %1 started at: >> l:\sn_commands\node%1.log
date /T >> l:\sn_commands\node%1.log
time /T >> l:\sn_commands\node%1.log
echo ---- --- -- - -- - >> l:\sn_commands\node%1.log
echo init > l:\sn_commands\job%1
goto start
:crashed
echo --= CRASHED NODE %1 =-- -- - >> l:\sn_commands\node%1.log
echo LWSN NODE %1 crashed at: >> l:\sn_commands\node%1.log
date /T >> l:\sn_commands\node%1.log
time /T >> l:\sn_commands\node%1.log
echo ---- --- -- - -- - >> l:\sn_commands\node%1.log
goto start
:start
@LWSN.exe -2 -dl:\ -cl:\sn_config l:\sn_commands\job%1 l:\sn_commands\ack%1
goto exit
:exit
if NOT %ERRORLEV% == 0 GOTO crashed
echo --= STOPPED NODE %1 =-- -- - >> l:\sn_commands\node%1.log
echo LWSN NODE %1 stopped at: >> l:\sn_commands\node%1.log
date /T >> l:\sn_commands\node%1.log
time /T >> l:\sn_commands\node%1.log
echo ---- --- -- - -- - >> l:\sn_commands\node%1.log