Queue is filled up but not getting processed

Use this forum to report bugs and to check for bugfixes and new releases of OpenEMM

Moderator: moderator

sp
Posts: 21
Joined: Mon Apr 12, 2010 2:04 pm

Queue is filled up but not getting processed

Post by sp »

The var/spool/QUEUE directory had over 10k files in it and didn't accept any new files:

Code: Select all

[26.07.2010  00:38:59] 3979 INFO/loop: Queue is already filled up
Only after a restart of OpenEMM things went on normally and mails from the queue were proccessed and new were accepted to the queue with the queue growing up to 15k.

As a workaround, which processes do need to be restarted periodically to avoid such a lock?

Is this a known issue? OpenEMM 6.0.1 is being used.
maschoff
Site Admin
Posts: 2597
Joined: Thu Aug 03, 2006 10:20 am
Location: Munich, Germany
Contact:

Post by maschoff »

A restart only leads to the unpacking of the next chunk of 5,000 mails, so it does not help really. You have to investigate why Sendmail can not empty its queue: Are receiving SMTP servers blocking mails? Too many invalid addresses or softbounces?
OpenEMM Maintainer
sp
Posts: 21
Joined: Mon Apr 12, 2010 2:04 pm

Post by sp »

ma wrote:A restart only leads to the unpacking of the next chunk of 5,000 mails, so it does not help really. You have to investigate why Sendmail can not empty its queue: Are receiving SMTP servers blocking mails? Too many invalid addresses or softbounces?
Thank you for your reply. We're not using sendmail but OpenEMM's semu.py (I guess it stands for Sendmail emulation :D).

I don't know why this happened. From what I understand there was no machine or network failure, so I suppose it was some kind of a software hiccup.

After the restart, the queue got down to 2k after accepting and delivering another chunks in a total of ~ 50k, so I am happy with the result, but we need to avoid such a lock where mail is not delivered and incoming new mailings are blocked by a filled up queue as recipients get easily upset when the recieve several mails in a short time span which were meant for daily delivery.

I suppose it is okay to restart "bounce.sh, merger.sh, mailer.sh, slrtscn.sh" and leave "core.sh" with Resin running? For a confirmation on this and any further hints I'd be very grateful.
maschoff
Site Admin
Posts: 2597
Joined: Thu Aug 03, 2006 10:20 am
Location: Munich, Germany
Contact:

Post by maschoff »

Yes, the backend scripts and the Java frontend work independently and communicate via the database.
OpenEMM Maintainer
sp
Posts: 21
Joined: Mon Apr 12, 2010 2:04 pm

Post by sp »

Thank you.
Post Reply