From: Green, Simon (SGreen@KRAFTEUROPE.COM)
Date: Fri Dec 13 2002 - 09:53:49 EST
So; just to make sure I understand the situation...
This is not a problem with a takeover as such; it's a problem getting the
cluster configured and running at all.
What versions of AIX and HACMP are you using? HACMP/ES or classic?
What's your experience with HACMP? (No criticism intended: I just want to
avoid telling you obvious stuff if you're used to working with HACMP.
Having said that, here are a few obvious things. :-))
If I were you, I'd go back to the beginning, (nearly!).
First of all, verify your cluster; both the topology and resource groups.
If that shows any problems then you need to sort them out before continuing.
Assuming it's all OK, I would then make sure that nothing starts
automatically: don't start cluster services automatically, or activate idle
takeover. Keep it all under your own control. I'd also keep the debug
level low; you just get swamped with messages otherwise. Once you've got
some idea of where things are going wrong then you can set it higher if need
be.
Modify your application scripts. Take out all the code which actually does
anything and just put in something like "touch /tmp/ha_application_startup"
so that you can see it's been invoked. This will allow you to play around
with stopping and starting the cluster and migrating resources without
having to worry about the application itself. If nothing else, it'll be
quicker!
Once you've got the basics working - resources acquired correctly - then you
can put back the application script and try it properly.
If you're still having trouble after this, some listings of the cluster
topology and resource group(s) would be helpful.
Simon Green
Philip Morris ITSC Europe
AIX-L Archive at http://marc.theaimsgroup.com/?l=aix-l&r=1&w=2
AIX FAQ at http://www.faqs.org/faqs/aix-faq/
N.B. Unsolicited email from vendors will seldom be appreciated.
-----Original Message-----
From: Patrick B. O'Brien [mailto:pobrien@DOIT.NV.GOV]
Sent: 12 December 2002 23:45
To: aix-l@Princeton.EDU
Subject: Re: Oracle and HACMP
It was complaining big time about one of my 1 Gig cards. Yet the wrap test
said all was good. I'm getting nailed with snmpd errors.
So, I'm not even out of the gates. My people are asking me if it is worth
it. I've been trying to get the test environment stable, with no errors and
I can not seem to get there. The client piece, clstat says the whole darn
thing is down.
I just configured 2 100 meg cards and I'm starting stuff up again.
I'll look close at HACMP.out.
When I talk to IBM if feel like I know as much as them.
-----Original Message-----
From: Ferenc Gyurcsan [mailto:fgyurcsa@AVAILANT.COM]
Sent: Thursday, December 12, 2002 12:35 PM
To: aix-l@Princeton.EDU
Subject: Re: Oracle and HACMP
What does hacmp.out have to say? How did you halt the node?
--Ferenc
-----Original Message-----
From: Patrick B. O'Brien [mailto:pobrien@DOIT.NV.GOV]
Sent: Thursday, December 12, 2002 1:46 PM
To: aix-l@Princeton.EDU
Subject: Oracle and HACMP
Has any one out there actually got HACMP to work with Oracle? Has any one
out there actually halted a box and the Oracle that is running on that box
starts on the other node in that cluster?
I'm having all kinds of HACMP problems and IBM does not seem to have the
answers.
Thanks.
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 22:16:24 EDT