Re: Handling SQLSTT=42703

[Buck Calabro <kc2hiz@xxxxxxxxx>]

On Mon, 12 Jun 2023 at 18:24, Reeve <rfritchman@xxxxxxxxx> wrote:

-snip-

> logic.  My coding standards don't allow "unmonitored" application
> error/exception messages, like SQL0206, in the job log.  The code in
> question is in a service program so the solution doesn't have to be
> duplicated in other source.

Consider RMVMSG for the SQL0206 that you are expecting.One way to do
this is to SNDMSG a 'marker' message, and then do RCVMSG backward to
the marker, inspecting the various messages in the joblog to see
whether you have anything you find to be a genuine error and RMVMSG
the ones you expect. The point being that there are times that we
expect things like CPF2105 for, say, a re-run scenario, and because we
consider these to be noise in the joblog, we can safely delete them.

I use this 'marker message' technique during my backup process to
inspect (but not delete) messages saying that this or that was in use
and not backed up. Some files, like MGTCOLsomethingorother are always
in use, but when the code comes across something unknown, it emails
me.

Perhaps this will spur other ideas from the list. Error
monitoring/handling is not often discussed here, and it's nice to see
someone else's thoughts on the matter.
  --buck