Sporadic yet frequent email issue

[lost108]

Hello,

We currently have about 200 domains on our server. Every so often, several domains will start bouncing back emails with the following error:

mail.info:Jul 14 08:26:39 srvrName postfix/virtual[20115]: 3DDA5E22FC8: to=<alex@aDomain.com>, relay=virtual, delay=0.06, delays=0.06/0/0/0, dsn=5.1.1, status=bounced (unknown user: "alex@aDomain.com")

Every single time, we check the database and see that the user in question does indeed exist. By changing the status of the domain's mail users to "change" and calling the ispCP request, the issue is immediately resolved until it occurs to another domain.

Has anyone encountered this same issue? What might be causing it? Is there a way to prevent it / manage it in real time so that our users do not miss any emails?

Thank you for your suggestions!

[Diggo]

Howdie Lost,

do you got any news on that issue? I'm actually having the same problem.

[lost108]

Hello,

No news, it continues to happen. I've also had the issue where the /etc/postfix/domains file loses nearly all of domains in the file. The only way to repair that was to set the domain's status to 'change' and have everything run again.

I see you're using RC6 so I'm assuming that issue hasn't been resolved in that release. Does anyone have any ideas that can help us determine the cause of this issue?

Thanks!

[Diggo]

lost108 Wrote:I see you're using RC6 so I'm assuming that issue hasn't been resolved in that release.

Hi lost,

i didn´t test it yet, but i think the error i get is caused by another problem (although it has the same error message...). I´ll shortly explain what my problem is:
I am using a different MX for some of my domains. The problem then is, that from the "outside" (from my isp-cp server) mailerdaemons look up the MX correctly in the DNS. From my server, for example generated password mails can´t go out. So anybody can mail me to diggo@domain.de but the server itself who carries my web on domain.de can´t. This is caused by postfix and/or the local DNS-MX entry. So postfix doesn´t even look up the MX for this domain, because ISP-CP generates a local delivery. The problem is, that the mailbox "diggo@domain.de" doesn´t exist locally, so the mail can´t be delivered and all you see is that error entry inside the syslogfile.
If you have a similar setup like me (some domains have a different mailserver), then it might be worth to take a look at the FAQ, where i recently added a entry to that topic.

Greetz!

[lost108]

I see what you're saying. I guess our issues our a little different. We have Bind disabled on our setup so that all DNS lookups must be made from the authoritative name servers.

If we want mail that was sent locally (i.e. via PHP) to be delivered to the correct MX server, we just remove the domain from the /etc/posfix/ispcp/domains file (don't forget to overwrite the /etc/ispcp/posfix/working/domains file as well). This will cause the local lookup to fail and an external lookup will return the correct MX server.

Not sure how it'd work with Bind running on your server though... Might still be a lost cause.

Hope that helps! Have a nice weekend.