[Feature request] introduce mailout check

Bug #968086 reported by Alexander Bittner
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
KARL3
Invalid
Medium
Alexander Bittner

Bug Description

Hello!

today, we received support requests from Ariadne and Oxfam. Both reported that mails sent from KARL did not arrive to the recipients (in this case: community invitations and blog post notifications)

We investigated and found the following in the /srv/multikarl/production/current/var/log/mailout.log:

2012-03-29 02:30:15,746 WARNING karlserve.scripts.main: mailout already running with pid 3246
2012-03-29 02:30:15,747 WARNING karlserve.scripts.main: Exiting.

It turned out that for some reason the lock file /srv/multikar/production/current/var/lock/mailout.pid had not been touched since quite a while and that meanwhile, there was a different process running with this PID.

Removing the PID file helped and mails that still were in the queue have been sent successfully.

To discover problems like these more quickly in the future, we wondered if it would be possible to introduce some kind of a mailout check in analogy to the one that checks for the correct mailin behaviour (mailin-trace).

Best regards,
Alex

Revision history for this message
Paul Everitt (paul-agendaless) wrote :

Sounds like a good idea.

Changed in karl3:
importance: Undecided → Medium
milestone: none → m98
Changed in karl3:
milestone: m98 → m100
Changed in karl3:
milestone: m100 → m104
Changed in karl3:
milestone: m104 → m108
Changed in karl3:
milestone: m108 → m112
Revision history for this message
Paul Everitt (paul-agendaless) wrote :

Chris, this looks like a good one to work on next week.

Changed in karl3:
assignee: nobody → Chris Rossi (chris-archimedeanco)
milestone: m112 → m110
Changed in karl3:
milestone: m110 → m111
Changed in karl3:
milestone: m111 → m113
Changed in karl3:
milestone: m113 → m115
Changed in karl3:
milestone: m115 → m116
Revision history for this message
Chris Rossi (chris-archimedeanco) wrote :

The circumstances that led to this are pretty rare and depend on a fairly remarkable coincidence.

As far as a mailout check, I'm not sure what form that would take. We could easily have a cronjob that caused an email to get sent through Karl's machinery to . . . . somewhere.

I guess there's a question for Gocept here. If we set up some kind of regular email, would you guys set up some kind of monitored endpoint that would trigger an alarm if the mails stopped coming?

Revision history for this message
Paul Everitt (paul-agendaless) wrote :

Added Alexander since Chris asked a question back to gocept.

Changed in karl3:
assignee: Chris Rossi (chris-archimedeanco) → Alexander Bittner (yb)
milestone: m116 → m117
Revision history for this message
Paul Everitt (paul-agendaless) wrote :

Alex, do you want to keep this one open?

Revision history for this message
Paul Everitt (paul-agendaless) wrote :

Per Theune, Alex is on vacation

Changed in karl3:
milestone: m117 → m118
Revision history for this message
Paul Everitt (paul-agendaless) wrote :

We'll consider this one dead.

Changed in karl3:
status: New → Invalid
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.