Topic: imap problems

[osc]

Hi,

i have some problems with my server.
After some time all imap/s clients (Mail, Outlook etc.) lose connection and get the error authentication failed.
I didn't find any erros in var/log/messages,  imap/current or /imaps/current.

Is there another logfile wehere i can find some hints on whats going on ?.

Thanks for your help.

Felix

[mmccarn]

Are these local IMAP clients, or remote?

Is the authentication failure occurring during email download (imap /imaps) or when sending  (qpsmtpd / sqpsmtpd)?

Does webmail work?

Is there something you do to make it work again (wait? reboot? restart imap or imaps?)?

Do you have any contribs installed that might affect server access (eg denyhosts)?

Is your server up-to-date?

[osc]

Thanks for the fast reply.

The clients are all remote.

Good that you mentioned webmail, when the clients lose connection webmail login is not longer working.

Restart the server is the only way to get it working again. I updatet the the server over the server manager.
It says up to date.

I reinstalled imp so webmail stopped working but thats an other problem...
No contribs installed that might affect the access.

[mmccarn]

You may be running into a "Path MTU Discovery" problem.

Check your MTU settings as described here for qpsmtpd errors:
http://wiki.contribs.org/Email#qpsmtpd_.22Connection_Timed_Out.22_errors

A problem with Path MTU Discovery does not generate any particularly meaningful errors anywhere...

[mmccarn]

...Also, (if your MTU is OK), I'd be curious if any of these get email working again:

signal-event email-update
(problem is in email services)

signal-event remoteaccess-update
(problem is in email services or firewall rules)

/etc/init.d/masq restart
(firewall became unstable for some reason - take a look at /var/log/iptables/current)

[edit: grammar, typo]

[osc]

MTU is 1500. tracepath resume with 1500.

I will try the other steps later.

thanks a lot.

[osc]

I tried signal-event email-update, signal-event remoteaccess-update and no errors in iptables.
I am still forcing the same problem.

I get my Webmail working again.
Thats the output from /var/log/messages.

[imp] FAILED LOGIN for username [x.x.x.x] to {localhost:143 [imap/notls]} [pid 7065 on line 139 of "/home/httpd/html/horde/imp/lib/Auth/imp.php

[mmccarn]

You might want to test your hardware - RAM, HDD, etc.

There's a RAM test on the SME server install CD.

[osc]

SME is running on ESXI 4.1.
I made a RAM test but no erros.
Other Hardware is running well.

[osc]

I tried telnet connection to imap.

Trying x.x.x.x...
Connected to x.x.x.x
Escape character is '^]'.
* OK imapfront ready. + stunnel
a1 login usernam password
Then it stucks.

After reboot its working and I can connect.

[mmccarn]

Does the clock drift on your SME server?  There used to be a problem with severe clock drift on virtualized SME servers (hours of drift in a few minutes).

[osc]

You are right.
I found in the IMAP Log.

imap(xxx): Fatal: Time just moved backwards by 97 seconds. This might cause a lot of problems, so I'll just kill myself now. http://wiki.dovecot.org/TimeMovedBackwards.

I have add the kernel parameter notsc divider=10 nohpet.
Hope that will fix it.

Thank you
Felix

[CharlieBrady]

Hope that will fix it.

Please edit the subject of this thread to include '[SOLVED]' if it does. Thanks.

[osc]

hi,
still the same problem.

[Stefano]

which one? time drifting?