AutoSys Workload Automation

Is anyone using Service Guard on remote agent machines?  We are having some intermittent issues on some new servers with Service Guard configured.  We found an old case way back in 1997 with AutoSys 3.3 that indicates there may have once been an issue and maybe there are some special instructions?

Hi Wade,

Could you please provide more details on the nature of the problem? What version of AutoSys remote agent are you referring to here? v4.5 or r11.0 SPx or the newer 11.3 SPx?

Is this the problem you are referring to on Remote Agent v3.3?

http://downloads.ca.com/akdlm/apars/d5/pATSYS/r3.3/oANY/PR461964.txt 

Thank  you,

Chandru

We are getting “end of file on communication channel” errors.

We are running R11.3.6 with system remote agent R11.4 SP4.

We are running to a 2-node Service Guard cluster of servers where they are active-passive in configuration.  We have set up the machine definition to point to the Service Guard “alias” and allowing that to direct traffic to the current active node.

"end of file on communication channel" errors as in "ORA-03113: end-of-file on communication channel"?

In which log(s)?

That seems odd in the context of agent clustering.

Does your machine definition looks something like this?:

insert_machine: <machine_name>

node_name: <service_guard_alias>

agent_name: WA_AGENT

port: 7520

If each agent has a different agent_name then it will be necessary to create an alias in each agentparm.txt.

If you define each machine explicitly do each successfully run jobs?

Regards,

Mark Hanson

The problem is a bit intermittent, but only experienced with setup of service guard machine as in your example. I am unsure of what you mean about alias in agentparm.txt. Will you please tell me more about that requirement?

Thanks,

Wade

Sent via the Samsung Galaxy S7, an AT&T 4G LTE smartphone

After looking at the alias parameter in detail I'm not sure it would apply in this case. It seems to me that in this scenario the agent_name's would all have to be identical in order to be addressed properly via the "cluster machine definition".

If your setup is working intermittently perhaps there is some external factor or AFM issue in play.

We would need to review logs for a root cause analysis.

On the ones where we have experienced the issue, any change to the shell script or reloading as a new job pointing to same script allows it to run fine.  Deleting and reloading same job again will usually cause failures as previously experienced.  Is there some sort of cache somewhere that we are not aware of or that comes into play when trying to execute through service guard?

There is IP caching on the scheduler side. This can be disabled by changing the following line in the $AUTOUSER/config.<ACE> file:

EnableIPCaching=1

to

EnableIPCaching=0

This might be the ticket!  We’ll check this out!  THANK YOU!