LoadBalance or Failover configuration for application server connectivity in Automic PeopleSoft Agen

[KedarKatabathula611016]

Hi All,

I am trying to see if anyone is able to create any loadbalance/failover application server connectivity for automic connectivity to PeopleSoft. I tested couple of things
1) Use a comma separated value and semicolon separated value, agent comes up but doesn't recognize the second part of the string.
2) I tried creating a load balance URL in our network to point to the two hosts at any time, the agent doesn't seem to come up.

any ideas are appreciated.

[Tran-Hoang_Hau_348]

Hi Kedar,

I would suggest to go with an easier, Like you can create 2 PeopleSoft Agents, each agent connect to a specific server, then put both Agents in one agent group.

http://docs.automic.com/documentation/AE/11.2/english/AE_WEBHELP/help.htm?product=awa#ucaclu.htm

Then when you run a job, it can run on any agent depend on the pre-set condition (like load balance, availability ..ect)

Hope that helps.

[KedarKatabathula611016]

Hi Tran,

Thanks for responding to the thread,

I could create multiple agents and create AgentGroups. The question still remains on high availability of PeopleSoft with Automic.

Lets suppose one of the application server goes down which is connected to one of the PeopleSoft Agent, I am thinking the related agent would still be up and running marking it as live in the AgentGroup. I will test it out and get back to the thread.

Also, each PeopleSoft Agent is a separate license which is not cheap.

[Anon Anon]

I think we have to seperate the Peoplesoft side and Automic side - from Automic side we only can improve the agent side.

So if Your Peoplesoft system goes down - no matter how many agents run-  automic has no influence on that and no possibility to recover it or react on it.

BTW: If that happens, the agent will keep running and the jobs will not get executed.

A far as I know (I am no peoplesoft Pro) some guys check the availability of the system by running a dummy job - if this runs fine the system is exspected as online and available)

Haus suggestion was for the scenario that the peoplesoft Agent is down.
The usage of 2 agents in an agent group is a common trick in productive important environments.

general hint:
you can check the availability (active/not active state) of an agent with script command: SYS_HOST_ALIVE in Prescript.

http://docs.automic.com/documentation/AE/11.2/english/AE_WEBHELP/help.htm?product=undefined#ucaapy.htm?Highlight=sys_host_alive

[Shane_Meek_183]

Hi Kedar,

How is the testing going? Will 2 PeopleSoft agents in an agent group satisfy the need here? As Wolfgang says, if the PS system or server goes down, our agent(s) will continue to run but jobs will not get executed or will error out. We have no governance on the status of the PS system itself. 

You can read more about the load balancing abilities of an agent group here.

Please let us know what other questions you have either here or via your case. 

[laura_albrecht_automic]

I don't have a PeopleSoft agent - but I do have a JDE agent and we have a similar situation here.  

We didn't want to use the agent group because then you lose the functionality of the form.  You CAN when you initially create the job point to a specific agent, configure it, then change it to point to the agent group.  That DOES work and takes care of the load balancing issue very nicely.  As long as you are OK with losing the functionality of the form this is fine.  You can still see the entries and even make changes to them - the drop downs just don't work because obviously it can't connect to an agent group to retrieve the data.

Anyway, the JDE guys here use a mechanism to determine if the JDE system is healthy by having some files on the server.  If these files exist - then JDE is good and we can run jobs there.  If the files aren't there - then we want to run on the 2nd node of the JDE instance.  They must have some kind of script written that handles this - I don't have all those details (I can get them if you want them), but it must periodically do some kind of check to make sure the JDE system is healthy - whatever that check is - and then remove the files if there is a problem.

On the Automic side I created an include which basically checks for these files and if they don't exist - move to the other server.

Again, not the same application, but just sharing something I thought might be applicable.