To All,
Over the last few months, OK years, we've all blasted OTS for not doing
their "job". For the most part it I think it is justified. Too many times the
"fix" is to upgrade to the latest software, which will take a number of days
which allows the analyst to close the TAR & get his/her brownie points. Well,
I've a case where I need to complement OTS. Yeah you got that right,
complement. Here is what happen:
On Wednesday night the good old pager went off at 8PM EST. Our locally
created Oracle monitor had a problem with the database that runs the current
generation assembly line. The error was "ORA-12500: TNS:listener failed to
start a dedicated server process"
I connected into our network (remote access is just the berries) and sure
enough I could not get into that server via any SQL*Net process. I logged on to
the computer (an HP L2000 running HP-UX 11.0) and queried the listener which
looked very normal. Also If I set the Oracle SID and used the BEQ protocol, no
problems. Humm, Since this machine does not use MTS I looked for the dedicated
server processes and found only 52 out of the "normal" 130. The listener.log
was no real help either, except for an "HP-UX error: 12: out of space" at the
end. Called the SA who likewise dialed in & could not see anything funny. Now
he & I are both stumped, so I tried restarting the listener and deleting the
listener.log. No help, OK time to call OTS. As normal they asked for the
alert.log and the newly created listener.log. The normal delay tactics, but
wait there's more!
Fifteen minutes after I finish uploading the specified files (MetaStink
being itself again) I get a return call for OTS. Your out of swap space they
say. Reboot the server & all should be well again. Yeah right! Pigs will fly
some day too!
So, I call the SA & inform him of Oracle's recommendation. This prompted
him to take a second look at the swap space settings & sure enough swap was less
than the HP recommendation of 200% of physical. Now since we're on a journaled
file system he could add swap space online & bingo, pigs start flying!
Why you ask, is this so? Well there is a "feature" in HP-UX that allocates
an amount of swap for every process in physical memory just in case the system
crashes so that a core dump exists. That's why they recommend that swap equal
200% of physical at least. Our real problem was a new process that spawned
copies of itself as required & the darn thing had gone ballistic eating up all
of swap. Both the SA & I forgot about this "feature".
So for once I need to get off of the rant box and pass a complement to one
group at Oracle that we all love to hate.
Dick Goulet
--
Please see the official ORACLE-L FAQ: http://www.orafaq.com
--
Author:
INET: [EMAIL PROTECTED]
Fat City Network Services -- (858) 538-5051 FAX: (858) 538-5051
San Diego, California -- Public Internet access / Mailing Lists
--------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from). You may
also send the HELP command for other information (like subscribing).
