RE: hp help...
- From: "Docherty, Heather" <H.Docherty@xxxxxxxxxxxx>
- To: "'oracle-l@xxxxxxxxxxxxx'" <oracle-l@xxxxxxxxxxxxx>
- Date: Tue, 20 Apr 2004 11:35:44 +0100
I had a few thoughts about how you can find out what happened on the
production box:
If the system itself went down, the system log /var/adm/syslog/OLDsyslog.log
should have entries for the period before the crash. Also, did the sys
admin have any of the HP monitoring software running, which is quite
comprehensive, and have they checked the relevant diagnostic logs and its
mails to the superuser?
I can't remember exactly, but I seem to recall that the install used to need
about 400MB free in /tmp. If you had the amount recommended free, then the
crash probably - definitely - or even possibly came about because the box
itself ground to a halt.
If so, I have seen this before and it looks to me as though the system may
have run out of processes it can allocate, and the sys admin needs to tweak
/ increase key HP-UX kernel parameters in line with the configuration of the
box i.e. number of cpus, memory etc. If swap is needed, the sys admin can
add this dynamically using sam, with no downtime needed.
I have 8i and 9i running on an HP L2000 - now an RP5470 - and had no trouble
installing 9i, although I did take a copy of the entire system and put it on
our Sap upgrade test box first to see what happened. (I suppose that I am
'lucky' in that I do the sys admin and oracle dba work, so I get to tweak
things to see what effect it will have.)
My advice in sys admin mode is: get the sys admin to contact HP as they
have a great paper on configuring HP-UX to run oracle, including algorothms
for correctly calculating the values for key kernel parameters relating to
the number of oracle instances running, number of users, recommendations as
to swap space size etc, get them to do the work - and then shoot them. I
recommend this course of action because I am presuming that you told them
what you were going to do, and they didn't act in response to your
notification. (I have been in this position myself, but had great
difficulty pointing the finger at me afterwards.............!!)
Personally, I would be interested to hear from anybody here with experience
of Sap, who may know of ways to reduce the current time taken (48 hours), to
update the tables PPOPD (@ 2.76 million rows) and PPOIX during a local
client copy.
Heather
-----Original Message-----
From: Chris Stephens [mailto:ChrisStephens@xxxxxxxxxx]
Sent: 16 April 2004 19:46
To: 'oracle-l@xxxxxxxxxxxxx'
Subject: RE: hp help...
flounder:/=>dmesg
Apr 16 13:28
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
Other related posts:
