Re: HA clusters and filesystem configuration for IPC

[KANE, TOM M (AIT)]

Michael,

 

One other approach that may not be recommended, but that does seem to work is to run a script to delete all the IPCS resources (owned by mqm) with the ipcrm command.  If you do this before MQ comes up then it will generally come up clean.

 

It's simpler than rebuilding your qmgr or moving stuff around.  And it seems to work.

Tom

-----Original Message-----
From: Michael F Murphy/AZ/US/MQSolutions [mailto:[EMAIL PROTECTED]]
Sent: Thursday, December 19, 2002 10:15 AM
To: [EMAIL PROTECTED]
Subject: Re: HA clusters and filesystem configuration for IPC

Thanks, Roland.  Great info.

There is a SupportPac for HP too but it doesn't look like it was used by my client because the moving of the ipc directories was not done.  I do think this is the problem.  Thanks for reminding me about the mounting over mounted filesystems.  That is the case here but I am fairly certain they are unmounting in the right order, but I should check just in case though.  As for the @SYSTEM, I had a problem with moving that one time but I don't remember what it was.  In that case, I was using HACMP.  In this current problem, I see FFST files created because resources are trying to access the @SYSTEM directory.  I think it is safer to leave it local.

Mike Murphy

Roland Duenki <[EMAIL PROTECTED]> wrote:

Date Recieved:		12/19/2002 08:47:30 AM
To:		[EMAIL PROTECTED]
cc:
Bcc
Subject:		Re: HA clusters and filesystem configuration for IPC

Hi Michael,

I have some AIX and Solaris HA systems...There are some supportpacks
with scripts for those two systems.
Maybe there are some for HP, too. Otherwise, I recommend you to read
the MC63 documentation...

Some things you should look at:
First of all, you should end the queuemanager before taking its
filesystems away. It will crash otherwise, leaving some nasty
errormessages behind.

Second, try not to mount one filesystem on another mountet filesystem;
you cannot unmount /var/mqm while /var/mqm/log is still mountet.
create /var/mqm/<xxx> and /var/mqm/log/<xxx> as mountpoints, and put
the qmgr in there. This also helps a lot when you have more than one
qmgr on the HA cluster.

Then about the IPC ressources:
As you mentioned, AIX and Solaris supportpacks have a tool named
"halinkmqm". It moves the IPC-directorys from your qmgr back to a
local disk (/var/mqm/ipc) and puts links to the qmgr directory. Those
files shall not be moved during takeover.
I had a lot of problems when I left the ipc directorys with the qmgrs.
All those problems where gone after using halinkmqm.

Finally: the @SYSTEM:
Actually, I always copy this to any /var/mqm/<xxx>/qmgrs directory.
Therefore it will move with the qmgr.
I dont know if this is good or bad. But it works.   :-)

Roland


___________________________________________________________________

Disclaimer:


Diese Mitteilung ist nur fuer die Empfaengerin / den Empfaenger
bestimmt.

Fuer den Fall, dass sie von nichtberechtigten Personen empfangen wird,
bitten wir diese hoeflich, die Mitteilung an die ZKB zurueckzusenden
und anschliessend die Mitteilung mit allen Anhaengen sowie allfaellige
Kopien zu vernichten bzw. zu loeschen. Der Gebrauch der Information
ist verboten.


This message is intended only for the named recipient and may contain
confidential or privileged information.

If you have received it in error, please advise the sender by return
e-mail and delete this message and any attachments. Any unauthorised
use or dissemination of this information is strictly prohibited.

"{--劉~ᯊjvx2 j)b b.ns+bvz'^v)ͅ²
ڕKnךi^jmᴶ %r쉀b᣽mᶟf᯷Iezwr