Jodie
05-18-2002, 12:17 AM
So - 1/3 of my farm crashed today. With statsman randomly down and/or doing some funny counting, it's hard to tell when.
Error logs down show anything other than the last start time.
These machines are rock-solid-stable being a renderfarm - which is atleast as tough as DF.
Anything happen today to kill 14 machines?
That's a worse record than G@H had with me - 1wk - 14 crashes. Ouch.
One more little thing, hopefully someone can point me in the right direction. I run on a cluster with the -qt option. When the client crashes, it runs through one sequence, uploads the data and then quits. Any way to get around that? Having to wait through a sequence and then start again is painful. I've made a second script - fcrash - that runs in interactive mode so I can atleast see when it stops after that first sequence, and then restart it.
Anything I should be concerned about?
Thanks!
--- Jodie
Error logs down show anything other than the last start time.
These machines are rock-solid-stable being a renderfarm - which is atleast as tough as DF.
Anything happen today to kill 14 machines?
That's a worse record than G@H had with me - 1wk - 14 crashes. Ouch.
One more little thing, hopefully someone can point me in the right direction. I run on a cluster with the -qt option. When the client crashes, it runs through one sequence, uploads the data and then quits. Any way to get around that? Having to wait through a sequence and then start again is painful. I've made a second script - fcrash - that runs in interactive mode so I can atleast see when it stops after that first sequence, and then restart it.
Anything I should be concerned about?
Thanks!
--- Jodie