07 July 2011, 14:20

Hello all,

I need a little direction. Our Mahara has recently started sending out old-e-mail messages -- and I don't know where to start looking to track down the problem.

We're running in a LAMP environment, and upgraded earlier this week from Mahara v1.1.6 to v1.4. No, I don't think that's what started the problem. The first re-send occurred before the upgrade and twice more since then.

I don't *think* the same messages are being sent out again and again. They tend to be fairly generic, without unique content so it's hard to tell; except for the Contact Us messages -- those were our first clue these were old messages. Also it's not the same number of messages each time.

I inherited the site from a previous admin and the user base is relatively small (~200), so the site has been suffering from benevolent neglect until recently.

Where do I start looking?
Where can I find the list of tasks triggered by cron?
Is there a database table of pending messages I can examine? (And clear out...)

Hopefully,

D. Hrynkiw
Kwantlen Polytechnic University

08 July 2011, 11:29

Could my problem be related to the topic in this thread?

Perhaps my users were mistaken when they reported that the first re-send occurred *before* the upgrade.

Any tips or hints on how to fix this -- or at least disable the flow of old e-mail for the time being -- would be much appreciated.

D. Hrynkiw

10 July 2011, 21:26

Donna, the flooding of emails on that thread are all due to a combination of a mail server misconfiguration, and a Mahara bug which sends way too many notifications to the site admin saying "User notification error was probably caused by your server configuration."

If you're getting that problem, it could be the same.  If you're getting old notifications being sent out, and they're all different, it's something else.

Have a look in the activity_queue table; it contains the notifications that haven't been sent out yet.  If there are rows in there that aren't being deleted after the notification has been sent out, ideally we'd like to find out why, but if you just want a quick solution, delete those rows.

11 July 2011, 15:04

Hello Richard, and thank-you so much for your reply.

My users are reporting that the messages are all different, so I don't think I have the same problem as the thread discussion.

My activity_queue table contains 19,000+ entries, dating back to April 2011 and appears to represent normal user activity. Your comment "it contains the notifications that haven't been sent out yet" implies to me that normally this table should be almost empty -- that is, containing only recent and unsent notifications, yes?

I need to give my poor users some relief, so I'll be deleting most of the rows.

Next step: why aren't the notifications being deleted after being sent out?

11 July 2011, 18:06

Hi Donna,

In terms of the notifications not being deleted, I suspect that you are affected by a bug (fixed in latest 1.3 and 1.4 releases) which kept notifications around when they failed, to be retried later.

If you are interested in a detailed analysis of why this happened, have a look at this post:

  http://feeding.cloud.geek.nz/2010/06/handling-email-based-events-safely.html

Cheers,

Francois