On Wed, May 08, 2013 at 01:30:45PM -0700, Brian McNally wrote:
Hi all,
qacct reports usage from a file, but GE has its own internal database
for tracking jobs and usage. Is this correct? If so, what controls
the length of time GE keeps job data for? It seems that using qacct
to display overall usage per user (-o), for example, might be a
little misleading if the actual accounting information is stored
internally. Users might draw conclusions about their usage and how
that'll impact their job priorities based on potentially incorrect
data.
I think--but do not know for certain--that the only internal historical
account info kept by *GE is for share tree computations. Functional
shares don't need it, since that's based only on a single point in time.
The share tree shouldn't need to keep information (in aggregated form)
for longer than the share tree decay time.
The accounting files can (and often do) grow without bound. I've seen
multi-GB files... Large files make for very slow response from qacct,
since it has to parse the entire file. Smaller files make for faster
qacct processing, so a periodic rotation is a good thing.
Thanks,
--
Brian McNally
_______________________________________________
users mailing list
[email protected]
https://gridengine.org/mailman/listinfo/users
--
Jesse Becker
NHGRI Linux support (Digicon Contractor)
_______________________________________________
users mailing list
[email protected]
https://gridengine.org/mailman/listinfo/users
