Re: HACMP automatic Oracle failover in non-RAC environment

From: Dan Norris <>
Date: Thu, 28 Aug 2008 07:49:42 -0500
Message-ID: <>


Yes, that's the easy workaround. I worked at a customer site with 6
nodes and about 50 crontab entries and at least one entry was changed
or added every few days. Copying it to 6 nodes was automated, but it
was still a manual action to start the "sync" to the other 5 nodes and
sometimes was forgotten with terrible results. That's why I'm a fan of
DBMS_SCHEDULER in failover clusters (apparently HACMP has shared
crontab, so I guess I'm talking more about those without shared crontab


Sais, Gene wrote:

Just copy the crontab over. It is located in /var/spool/cron/crontabs.


From: [] On Behalf Of Fuad Arshad
Sent: Wednesday, August 27, 2008 11:22 PM
Cc: oracle-l-freelists
Subject: Re: HACMP automatic Oracle failover in non-RAC environment



Our HACMP clusters the crontabs and software  is shared  which basically helps it failover with the system.
This eliminates the hassle of worrying about 2 crontabs. As far as backups go legato has pretty good API to manage HACMP nodes for non RAC system.

----- Original Message ----
From: Dan Norris <>
Cc: oracle-l-freelists <>
Sent: Wednesday, August 27, 2008 8:51:29 PM
Subject: Re: HACMP automatic Oracle failover in non-RAC environment

I've been around lots of failover clusters like these and one of the challenges that I've seen in these environments is scheduling routine scheduled tasks. Often, crontab is used to start backups and in a failover cluster (especially when all nodes are treated equally instead of having a "primary" and "backup" node) and this poses a problem since crontab isn't "cluster-aware". This is a great time to get familiar with the DBMS_SCHEDULER interface that allows you to kick off shell scripts. That way, you'll always ensure that the backup script will run on the same node where the DB instance runs since the instance starts its own backup.

Other than that little tidbit, there shouldn't be anything else that would be dramatically different. Just make sure that you put all the pieces of the database on the proper storage locations (locations that will either be failed over or that are available to all nodes). I'd also recommend scheduling a failover test at least a couple times every year just to make sure that you aren't surprised by a non-starting instance during a failover time.



Is anyone currently doing or have any thoughts or comments

on using AIX HACMP automatically to fail-over Oracle in a non-RAC



Do I need to worry about data corruption or anything else?  We have

manually done this and I've restarted Oracle, shutdown cleanly

and then restarted again.  Done my sanity checks and all has been



I've been ask about automating this failover w/scripts.  My thoughts are

I need to control this process and be aware but I also realize

hardware and software are becoming more and more sophisticated.




Please consider the environment before printing this e-mail

This e-mail, facsimile, or letter and any files or attachments transmitted with it contains
information that is confidential and privileged. This information is intended only for the use of the
individual(s) and entity(ies) to whom it is addressed. If you are the intended recipient, further
disclosures are prohibited without proper authorization. If you are not the intended recipient, any
disclosure, copying, printing, or use of this information is strictly prohibited and possibly a
violation of federal or state law and regulations. If you have received this information in error,
please notify Children's Medical Center Dallas immediately at 214-456-4444 or via e-mail at Children's Medical Center Dallas and its affiliates hereby claim all
applicable privileges related to this information.


-- Received on Thu Aug 28 2008 - 07:49:42 CDT

Original text of this message