Actually, the difference is very real, and you will probably be unable to run
Oracle without making changes. The speed of your processor or the amount of RAM
you have don't matter much here. Increasing kernel parameters are necessary to
meet the System V IPC requirements that Oracle makes in order to run. If you
make heavy use of Oracle--large tables, heavy usage, etc. you will probably find
the need to increase some of these parameters even further. It all depends on
how Oracle was written, how it handles IPC shortages, and what other programs
you may be running that have IPC requirements.
Don't think of this as "hacking" your kernel. It's very common for UNIX
sysadmins to have to tune kernel parameters to suit their users and the
applications they run. I work on software which runs across lots of UNIX
platforms (Solaris, AIX, HP-UX, Digital UNIX and others--Linux coming soon) and
we frequently tell people to bump up kernel parameters for System V IPC. In my
experience, AIX is the only UNIX which avoids all of this (SEM*, SHM*, MSG*) by
handling it dynamically and transparently. The only quirk is that shared memory
on AIX behaves a bit differently than on other UNIX platforms (look up the
EXTSHM=ON environment variable on the web and you'll see what I mean).
While we're on the topic of IPC, does anybody have information on the state of
POSIX.1b IPC (a la Stevens' UNPv2) in Linux? I did a bit of looking around just
now and the best piece of information I could find was an old (1996) l-k note by
Ulrich Drepper. Will there be a new set of tuning parameters to deal with or is
that going to remain a "feature" of System V IPC only?
Cheers,
Justin T. Fries
MQSeries Support
Raleigh, North Carolina
"Just three miles west of the Angus Barn"
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/