Subscription emails from Rob's blog not arriving.

[sam]

A user has got in touch to let us know he is not getting Rob's updates by email. He's still subscribed.

I just checked my mail and the last one I received was on the 11/4. I'll have a look and see if I can work it out.

[sam]

Using the test function seems to generate the content correctly: ​https://www.transitionnetwork.org/admin/messaging/notifications/events/test

[chris]

If you can send me the email addresses that haven't had the notifications off-list then I can grep the maillogs for these addresses to see if I can find out where the problem is -- it could be at a Drupal level, or SMTP level.

[sam]

Thanks Chris. Email sent..

[chris]

Looking at the two email addresses, the first one has a couple of recent log entries that I expect relate to the problem being reported using a form on the web site and a reply being sent.

Prior to that emails were sent to that address on:

• Apr 11 00:56:12
• Apr 11 00:56:11
• Apr 9 22:57:09
• Apr 8 20:58:53
• Apr 8 20:58:50
• Apr 4 20:56:18
• Apr 3 18:56:32
• Apr 2 16:57:21
• Apr 1 16:56:31
• Apr 1 16:56:30
• Apr 1 16:56:30

There have been none since April 11th. There are no errors in the logs.

For the other address emails were sent on:

• Apr 14 10:22:24
• Apr 11 12:59:48
• Apr 10 18:59:58
• Apr 10 10:59:51
• Apr 8 18:58:40
• Apr 8 10:57:20
• Apr 4 08:57:15
• Apr 4 08:57:14
• Apr 3 10:56:37
• Apr 2 08:57:18

Again no errors -- it very much looks like a Drupal issue and not a SMTP issue.

If we assume that email, on the 14th was for something else then it looks like the last notification was sent on 11th April.

I can't see what might have changed on 12th April from the ​https://tech.transitionnetwork.org/trac/timeline but this doesn't include comments, it would be handy if it did... perhaps that is something we can change.

The last BOA upgrade was started at Fri Apr 11 21:55:51, see ticket:707#comment:19 so I guess that rules it out as a cause since the last email appears to have been sent after that.

I have assigned the component for this ticket to "Drupal modules & settings" and have added Jim as a CC as he might instantly realise what the problem is.

[chris]

Following the suggestion here, ​http://www.gossamer-threads.com/lists/trac/users/40082 I have enabled an extra tick box on the timeline so we can see comments if we need to.

• ​https://tech.transitionnetwork.org/trac/timeline

The blog email notifications stopping appear to only coincide with the BOA update and the Heartbleed updates, the first blog entry that the reporter of the bug didn't get appears to have been this one:

• ​https://www.transitionnetwork.org/blogs/rob-hopkins/2014-04/impact-were-having-zsanett-and-zolt-n-transition-hossz-het-ny from looking at the RSS feed ​https://www.transitionnetwork.org/blogs/feed/rob-hopkins, it was published at 06:37:52 +0000

However I'm afraid that I'm at a loss to know how those updates could have broken Drupal email notifications.

[jim]

Cron hasn't run for 2 weeks 3 days according to ​https://www.transitionnetwork.org/admin/reports/status

So the upgrade didn't complete and restart the site crons... Was Octopus updated straight after Barracuda?

The hosting queues aren't running I'd guess (can't check from work) which would explain the lack of Cron.

[jim]

FYI I just ran cron.

[chris]

Replying to jim:

Was Octopus updated straight after Barracuda?

I'm really sorry but I'm afraid it appears that I forgot to do that. Is it safe to do it during the day or is it best done late at night / at the weekend when there are less people using the site?

There is an outstanding BOA update that hasn't been applied, BOA-2.2.3, this hasn't been applied due to the maintenance budget running out this month, see ticket:721

Would running it re-enable the Drupal cron task?

Following is the current root crontab:

* * * * * bash /var/xdrago/second.sh >/dev/null 2>&1
* * * * * bash /var/xdrago/minute.sh >/dev/null 2>&1
* * * * * bash /var/xdrago/runner.sh >/dev/null 2>&1
*/5 * * * * bash /var/xdrago/manage_ltd_users.sh >/dev/null 2>&1
03 * * * * bash /var/xdrago/clear.sh >/dev/null 2>&1
01 0 * * * bash /var/xdrago/graceful.sh >/dev/null 2>&1
08 0 * * * bash /var/xdrago/mysql_backup.sh >/dev/null 2>&1
58 2 * * * bash /var/xdrago/daily.sh >/dev/null 2>&1
58 5 * * * bash /var/xdrago/fetch-remote.sh >/dev/null 2>&1
###EOF2014###

This is the aegir crontab:

SHELL=/bin/bash
PATH=/opt/php53/bin:/sbin:/bin:/usr/sbin:/usr/bin
*/1 * * * * /var/aegir/drush/drush.php '@hostmaster' hosting-dispatch >/dev/null 2>&1

I can't find any others.

[chris]

Unless I hear otherwise I'll do the BOA-2.2.3 upgrade tonight, ticket:721 and hopefully this will resolve this issue.

[ed]

OK!

[chris]

According to ​https://www.transitionnetwork.org/admin/reports/status cron is running again, there has been more mailserver activity:

• ​https://penguin.transitionnetwork.org/munin/transitionnetwork.org/puffin.transitionnetwork.org/postfix_mailvolume.html
• ​https://penguin.transitionnetwork.org/munin/transitionnetwork.org/puffin.transitionnetwork.org/postfix_mailqueue.html

So I think this ticket is ready to close if someone can confirm they are getting the emails again.

[sam]

Hi Chris

Thanks for this. I can confirm the emails are now coming so will close the ticket.

The backlog all arrived in peoples inboxes in a mad flurry, so if this happens again I reckon we should delete the queue before restarting Cron..

Thanks

Sam

[ed]

Absolutely. Sam I'm sure you've got all the complaints under control (ie I've seen some but am not responding them so we dont' cross wires)

[chris]

Replying to sam:

The backlog all arrived in peoples inboxes in a mad flurry

Oh no, once again very sorry about this :-(

so if this happens again I reckon we should delete the queue before restarting Cron..

Good idea, but I don't know how to do that -- the queue was at a Drupal not a mailserver level.

[paul]

Delivery of subscription emails can be managed at admin/messaging/subscriptions