Define agents on Failover/Secondary server

rombapa · Post by **rombapa** » 03 Jan 2012 12:28

Hi,

I am in the process of preparing the failover system and I'd like to know if it's possible to define agents on the Failover server while the latter is down.
The "New Control-M agent" option is greyed out in the Configuration manager but I am wondering whether this can be done on the command line with some utility.

Thanks in advance

pmdeshayes · Post by **pmdeshayes** » 09 Jan 2012 3:11

You don't have to define Control-M/Agents on failover Control-M/Servers since the database replication does this job for you (if you have a recent version of Control-M of course).

All you have to do is make sure that both your Control-M/Servers hostnames are authorized on each of your Control-M/Agent.

Cheers

rombapa · Post by **rombapa** » 09 Jan 2012 4:05

Thanks pmdeshayes.

I am using Control-M 6.4.01.

The administration guide ("Preparing the failover CONTROL-M/Server" section) says:

The failover CONTROL-M/Server should be identical to the primary
CONTROL-M/Server and use the same CONTROL-M/Agents and remote hosts.

Doesn't that mean I should define the agents beforehand ?

Also: by "make sure that both your Control-M/Servers hostnames are authorized on each of your Control-M/Agent", do you mean the names of the servers should appear in the /etc/hosts of the Agents?

Thanks for your time.

pmdeshayes · Post by **pmdeshayes** » 09 Jan 2012 6:08

Failover relies on database replication and Control-M/Agent definitions and status are stored in its database. The idea is to be able to lose a minimum of data and time in case your Control-M/Server (A) crashes.

In a failover configuration, when you add a Control-M/Agent on Control-M/Server (A), the Control-M/Server (A) writes this information in its own database (A) and replicates on database (B).
As a consequence when you switch from Control-M/Server (A) to Control-M/Server (B) via the menu (guessing you are using BMC standard failover system), your Control-M/Agent is already declared and jobs can be submitted by Control-M/Server (B).

Should you want to check this out for yourself, use the following query on both your databases:

SELECT nodeid, hostname, domain_name, version, platform, os_name
FROM cmr_nodes
ORDER BY 1;

Note: This query is for Oracle and should work as is with PGSQL and Sybase (just replace ';' with 'go')

This query shall return the same result if your replication is correctly done.

As for the configuration on the Control-M/Agent side, you have to make sure your CONFIG.dat file have the following entries:

CTMSHOST <ctm_server_A>
CTMPERMHOSTS <ctm_server_A>|<ctm_server_B>

That is of course something you do in a "classic" failover system. In other configurations using VCS, alias/IP failover, Multi-tier replication (Dataguard/RAC/...) your settings will differ but that's another story and another level.

Cheers

rombapa · Post by **rombapa** » 10 Jan 2012 12:16

I get it. Thank you very much for your clear explanation.