[torqueusers] reading the "Time Use" in qstat

David Golden dgolden at cp.dias.ie
Fri Sep 8 05:18:14 MDT 2006


On Thursday 07 September 2006 19:57, Garrick Staples wrote:

> 'qstat' has always had CPU time.  'qstat -a' has always had walltime.
>
> > Might be time for me to read the qstat man page again... :-)
>
> The manpage does detail each column in the various display modes.  The
> standard mode is specified in POSIX and SuS, so we can't really change it.

I think what I was trying to remember was that the behaviour changes depending 
on whether or not cput was specified, the same output column in some cases 
(like -a) can end up with either cput or walltime in in it on a per-job 
basis.  That is precisely the documented behaviour in the man page,
could trip you up if you're naively awking output for some reason (... umm,
without reading the man page...)

I'm not sure that torque/openpbs always complied with that, but cput 
specification wasn't always mandatory on our system, so at some 
point, "qstat -a" (not "qstat" though I guess) would have indeed
started acting different.  So I'm not totally crazy. :-)

qstat -a , no cput in job (except used cput) or queue, it's definitely 
reporting walltime specified and elapsed, this just an interactive job 
sitting at a shell prompt for 5 mins

qstat -a:
58375.myhost  dgolden  Xn4t720f STDIN   7673  4  --    -- 07:00 R 00:05

However, interactive job sitting at a shell prompt with a cput specified for
5 mins:

qstat -a:
58376.myhost   dgolden  Xn4t720f STDIN  7756  4  --    --  28:00 R 00:00



More information about the torqueusers mailing list