Mainframe, MVS and zOS Discussion

Hi,

We are using REXEC on MVS server to execute perl on aix side . But we saw an issue recently : Perl completed successful at aix side , but the batch job still shows 'running' status . And after 4 hours and 40 minutes ( exactly 4 hours and 40 minutes) , the batch job will abend with reasons that didn't received correct completion code from AIX side .

And this issue only occurs for big jobs that runs more then 1 hour . Small jobs using same method works fine.

I need your help to decide what might caused the issue ? Is it AIX issue ? or MVS issue ? or Network issue ?

Thanks.

These issues are tough to get solution on a Forum, however, how do you make sure that the Perl was successful at AIX box?

This is one of those issues where the problem could be on the AIX component (such as the Perl script not providing the correct termination information to be passed back to z/OS), the network component (dropping the connection for various reasons, or mis-configured router / switch, etc), or the z/OS component (TCP/IP issues, for example).

Your ONLY way to resolve this problem is to work with your site support group. The 4 hours and 40 minutes sounds like a TCP/IP time setting, but your site support group would need to confirm that. You may need to have the site support group do some communications tracing to see what is going across the network (but if there's a test system available with a shorter time out than 4 hours 40 minutes, you'd want to use it -- there can be a LOT of communications traffic on a line in that amount of time).