Hi
We came across a bit of a strange problem yesterday where we had spare
cpus available for jobs, but there were certain ones not picking up
frames. We'd seen it before, but were never really sure whether it was
down to the way people had set their jobs up or something else. However
this time I was able to check all the jobs were set up correctly & I
also found these kinds of messages in the log on the job server for
these jobs:-
ALERT Ignoring frmarb 'Run': task in unexpected state 'Idle'
(expected Start|Busy) msg from ?@lafarm23:33523 * Lots of these showing
up for different farm machines
FAIL/LISTCPUS Fputs[2]: write failed: _SureWrite(): Broken pipe
ALERT Ignoring 'Idle': task in non-applicable state 'Start' for jobid
lin2.928 from ?@lafarm19:
lafarm19 & lafarm23 were two of the machines not picking up frames.
Also I've just checked through the logs this morning & found quite a few
of these types of messages on that job server:-
ALERT Task 'CpuPass1' ignored for non-existant frame -99999 from
?@lafarm23:33566
Prev=lin2 0 lin2.892,091_070_tiles_v04 -99999
100 2048 JobPass Job state is 'Done'
New=lin2 0 lin2.892,091_070_tiles_v04 -99999 100
2048 CpuPass2 Ram unavailable on lafarm23 (2048>0)
These only appeared between 4 & 4:15 am, and I'm fairly sure no one was
here rendering then...
Any ideas?
Cheers
Andrew
|