every now and then (one or two times in a year) the emails of a newsletter sent out by collective.dancing do not get delivered because they got stuck in the mailqueue of the MailHost.
~/buildout/var/mailqueue/new$ ls -l | wc -l
singing&dancing uses the default MailHost which is configured with "Use mail queue"
the mails can be sent out by restarting the queue processor thread in Plone/MailHost/manage (Plone/MailHost/manage_restartQueueThread?action=start)
the documentation sates, that the queue is automatically started when sending an email
however, this seems not to be the case when sending out mails singing&dancing.
def __init__(self, message):
self.message = message
msg = self.message
delivery = component.getUtility(zope.sendmail.interfaces.IMailDelivery)
delivery.send(msg['From'], self._split(msg['To']), msg.as_string())
except Exception, e:
# TODO: log
return u'error', traceback.format_exc(e)
return u'sent', None
>>> split = Dispatch._split
maybe anybody here experienced the same and has some hints what's causing these sporadic failures.
@keul: you mentioned something similar in Mailhost queues
(running plone 4.3.11)
Maybe simplistic but
mailq -q supposedly attempts to deliver all queued mail (I've used that in the past when I noticed a big queue). Is there another underlying problem, e.g. disk space? Throttling?
Maybe use a service like mailgun?
disk-space is no problem - the machine has several gigabytes free on all partitions and also gets monitored by icinga.
i think you got me wrong: it's not that the machine can't send these emails out (postfix is fine and fully functional)
the mailqueue built into MailHost fails to send out queued emails because the thread that is supposed to do this is not running.
I spotted this bug:
The mail queue might get stuck · Issue #45 · zopefoundation/zope.sendmail · GitHub which might be related to this because it causes the mail queue to hang.