[SOLVED] Lost messages: How to debug to find the cause of the problem?

[angelrengifo]

Hi:

I'm the administrator of a ZCS OSE 6.0.x integrated with MailScanner running on RHEL 5.4 x86_64. The system is working fine but several users are reporting that some sent messages never arrive to the inbox of the destination user.

According to that I asked for info of those messages like:

- Sender address
- Destination address
- Date and time

With this informationn I proceed to look at /var/log/maillog and I find that messages that was successfully delivered according to postfix logs like:

relay=mail.domain.com[172.27.2.2]:7025, delay=3.9, delays=3.5/0/0/0.43, dsn=2.1.5, status=sent (250 2.1.5 Delivery OK)

This messages just tells me that Postfix (Zimbra's MTA) delivered correctly the message to the Zimbra's POP3/IMAP server (vía LMTP?), right?
So I asume that this message wasn't deleted/retained/quarantined/etc... by MailScanner nor Postfix.

Then I try to look for the messages at the mailbox folder of the user (at /opt/zimbra/store/...) but I don't find the message because the user already downloaded and deleted it vía their MUA (MS Outlook) using POP3.

That's all I can do to "probe" that messages aren't lost. But users still keep claming that some messages get lost (occurs randomly).

I suggested to the local administrator of this company that verify things such as:

- Junk folder of the MUA (MS Outlook)
- Custom rules/filters of the MUA (MS Outlook)
- Antivirus quarantine & report
- Even suggested to stop using the MUA and just use Zimbra Web (no success, users can't stop using their favorite crap called Outlook)

So now I'm not sure who's responsible of this lost messages. Is it Zimbra? Is it the MUA? Is it the network causing errors on communication between client and server?

I would like to know if there's a way to turn on a debugging mode of Zimbra of the POP3 server, so I can know what messages got downloaded by whom.
It would be great if I can have a log of every message stored at POP3 server of Zimbra and when was downloaded via POP3.

I hope someone can help me with some ideas.

Thanks

[phoenix]

Originally Posted by angelrengifo

I suggested to the local administrator of this company that verify things such as:

- Junk folder of the MUA (MS Outlook)
- Custom rules/filters of the MUA (MS Outlook)
- Antivirus quarantine & report
- Even suggested to stop using the MUA and just use Zimbra Web (no success, users can't stop using their favorite crap called Outlook)

You should also check that there are no rules set in the Zimbra Web UI for these users. You could also stop them from using POP3 and get them to use IMAP (or even Zimbra Desktop). I'd take this as an opportunity to force a change of habits on the users.

Originally Posted by angelrengifo

I would like to know if there's a way to turn on a debugging mode of Zimbra of the POP3 server, so I can know what messages got downloaded by whom.
It would be great if I can have a log of every message stored at POP3 server of Zimbra and when was downloaded via POP3.

Have a read of these articles:

Ajcody-Logging - Zimbra :: Wiki
Using log4j to Configure mailboxd Logging - Zimbra :: Wiki

[angelrengifo]

Thanks, your suggestions helped me to find the cause of the problem trough /opt/zimbra/log/mailbox.log messages.

Finally, there were a lot of messages that Zimbra just deleted because of duplicated Message-ID values.
So I decided to change disable this feature by setting:

$ zmprov mcf zimbraMessageIdDedupeCacheSize 0

Now there aren't any lost messages, even if in some cases some messages arrive with duplicated Message-ID.

Thanks