SMTP Crashing Issues

Yolanda Harvey · Posted: Fri Jul 20, 2018 12:53 am

There is an issue with our mainframe SMTP. Daily we send out a message to certain users on our email system (Outlook). Back in June, there was a problem with the daily email, because one of the ids on the email was no longer valid, the email id was removed, but last week, somewhere in the pipeline between SMTP and Email, those old messages started being looped back to the SMTP ID I set up for BADSPOOLFILEID, and in the process flooding the application. Those old emails are looping back to SMTP over and over, which is causing it to crash and restart.

I removed SMTP from the autolog, so it no longer restarts, but have had no luck in figuring out why the messages are being looped back to SMTP. I have the email team looking at the messages, but am wondering if there is anything I can do from a mainframe perspective. We have IPLed the system, but that did not clear up the situation. I've tried getting answers from IBM, but unfortunately our zOS is out of support, so getting their help is basically not happening. Does anyone have any ideas that could point me in the right direction?

Robert Sample · Posted: Fri Jul 20, 2018 1:17 am

I would look in the SMTP configuration parameters (usually in TCPPARMS) at MAILFILEDSPREFIX and see if there are any data sets starting with that prefix. If so, and there is an ADDRESS and a NOTE data set (for each email there should be one of each), that could cause your problem. Delete the ADDRESS and NOTE data sets that exist and restart SMTP. These data sets are (usually) temporary and exist as a way for SMTP to keep track of undelivered messages; they generally go away as soon as the email is sent to the mail server.

Yolanda Harvey · Posted: Fri Jul 20, 2018 1:46 am

I think I did just fix it, and what you suggested is what I tried right before I saw your reply. I deleted all the ADDRBLOK and NOTE datasets (there were literally thousands) and it seems to be running okay now.

Thanks for the quick reply.

Robert Sample · Posted: Fri Jul 20, 2018 2:42 am

If the ADDRBLOK and NOTE datasets are hanging around, my experience is that usually a bad email has plugged the SMTP connection (yeah, that's not exactly what happens but it's understandable). Another thing to check is the SMTP log in the SMTP started task; the log is a good thing to have available since it gives information about each email.

Robert Sample · Posted: Fri Jul 20, 2018 3:11 am

Update: I found www.ibm.com/support/knowledgecenter/en/SSLTBW_2.2.0/com.ibm.zos.v2r2.hald001/addrbk.htm which says

Yolanda Harvey · Posted: Thu Aug 23, 2018 3:31 am

As I stated last month I deleted the ADDRBLOK and NOTE datasets which cleared up the crashing issues of SMTP, but what has happened since is that after a couple of weeks those datasets build up again, which causes SMTP to get really large in size, which would lead to it crashing if I were not monitoring it daily. The SMTP system is set up to send out a report daily, but if one of the people on the email chain has auto-reply enabled in Outlook, then that auto-reply kicks back and then gets caught in a continuous loop in which an autoreply response is sent back to SMTP over and over.

My question is this, is there a parameter that would disable SMTP from being able to receive auto-reply responses? Thanks.