How to setup ControlM Shouts from scratch??

Post Reply
sysad4

How to setup ControlM Shouts from scratch??

Post by sysad4 » 15 Oct 2009 8:56

I am a newbie using ControlM. I want to manage my jobs such a way that when a jobs fails ControlM will send out an email notifying me that a job ended not ok.

I want to know to set this up from scratch.

I have read Chapter 2: Utilities of the Control-M/Server UNIX Adminstrator Guide. But the explanations in the manual are not too high-level in approach for newbies like me.

Follow up questions:

1. the ctmshout utility requires an orderid of a job. Does this mean that every new day refresh, i need to execute the ctmshout utility and input all the orderid's of the jobs? Is there an easy solution for this?

2. In the Shout Destination Table, what do i need to input in the Addr field? is it S or A? and why?

3. according to the manual a logical name is a name used in the Shout or Do Shout parameter if the job processing definition to identify the recipient of the Shout message. So what does this mean? Is the logical name the email address of the recipient of the alert?

4. for physical name, can I just input any email address i want that the alert will be sent to?

Once everything has been setup, do i need to activate some setting in order for the email alerts to start working?



TIA

User avatar
markf
Nouveau
Nouveau
Posts: 134
Joined: 13 Jul 2006 12:00
Location: Switzerland
Contact:

Post by markf » 15 Oct 2009 10:10

ctmshout was one of the first 'great' things about Control-M, hence it looks a little strange at first. The best way to think of it is as an address book that maps to email addresses and that, when you are more familiar with Control-M, can be used for more complicated messaging functions.

Points -

1. No, you have misunderstood - the ctmshout utility is only used when you want to do specific shouts, most times you won't be using it as a command and you can ignore it at the moment. There is no daily housekeeping related to shouts.

2. 99% of the times this will be 'S' - the agent option is only if you want to generate the shout from the Agent platform (i've never seen anybody do it as 'A' but there's no reason why it wouldn't work so long as you have basic mail infrastucture on the Agent box). Just leave it as 'S'.

3. This is just a simple mapping. Enter the shout table details like this -

Destination = M
Adr = S
Logical Name = j_smith
Physical Name = john.smith@ibm.com

John Smith at IBM can expect an email whenever you use j_smith as a destination in your shout messages.

4. Yes.

There was a time when this was the only way to configure things. However we now have the Do Email option which does not need any of the above configuring (you just have to add the details each time you use Do Email in the Steps panel). However, most people prefer to stick with a shout table as it is fairly easy to administer and it also gives the possibility of 'shouting' to programs and passing job related variables through.

If you using just the default values and just the one shout table (called SYSTEM by default) then you don't need to turn any switches on to get the shouts to work.

Post Reply