CLEANMAIL – the folly of assumption

Bit of a short one this month as preparing for Year End tasks and shutdowns around the customers, but it gave me a little chuckle when I did it so maybe I can warn some people off the same fate!

CLEANMAIL, if you aren’t familiar, is a utility from IBM that’s useful for removing old distribution entries for users. You can find the source here and if you read the article by Graham Street last month then good news, it comes as part of the QMGTOOLS library so you’re already half way there.

For the purposes of the exercise I’m going to go with having manually created the CLs from the IBM page as opposed to using the downloaded version.

On a dark and stormy day in September last year we found out our message monitoring solution wasn’t working, my first thought was it’s a shame I don’t get email notification that my email notification system isn’t working! A bit of digging identified that the distribution limit had been reached so naturally I thought “I know the solution, I’ll run the mail cleanup utilities.” After it had run over all users, I restarted the subsystem and to my surprise, still didn’t work.

That’s very odd says I, so I check my notes in our knowledge base and confirm that the date there is anything prior to 2076 (same as the IBM sample code.) I don’t have the source on this customer server but I’m sure it must be right…..

As you may have guessed from that previous statement, it really would help if I had the source available. I go back to one of our internal servers that I saved the objects from, lo and behold someone had set a more sensible date in the program. So, the reason my purge wasn’t working is because by that date there was nothing to purge! The thought process at the time was “lets keep <x> months of entries” which is fine assuming you remember that / have a copy of your code next time the problem pops up.

Red faced after the amount of time it took me to get to this point, I sheepishly copy the source to the customer system and update the cut-off date. Would you believe this happened to fix my problem.

I still need to solve why my monitor is hitting this problem (he says writing about it 15 months later, how the time flies!) But hopefully if I don’t solve it by the next time the problem comes around, I’ll at least remember to check my source first 🙂