Jim Perrin jperrin@gmail.com wrote: On 3/28/06, spart cus wrote:

Hi To All Gurus, Im encountering some problems on my mailserver. When my clients are downloading their emails, when they reach around 15 mails downloaded somehow the connection stops. Im using Centos4.2, Postfix and Dovecot. Any one experiece these ?

So, not to ask the obvious question here, but what do the logs say? org/mailman/listinfo/centos Logs for my dovecot and postfix seems to be ok. Sending is not a problem so it means my Postfix works well. My receiving services is the problem. Cant find any error logs on my mail server.

--------------------------------- New Yahoo! Messenger with Voice. Call regular phones from your PC for low, low rates.

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

On Tue, Mar 28, 2006 at 10:01:06PM -0500, Jim Perrin wrote:

...

So, not to ask the obvious question here, but what do the logs say? org/mailman/listinfo/centos Logs for my dovecot and postfix seems to be ok. Sending is not a problem so it means my Postfix works well. My receiving services is the problem. Cant find any error logs on my mail server.

Iptables rules getting in the way? anything unusual in /var/log/messages?

Or on dmesg ?

- -- Rodrigo Barbosa rodrigob@suespammers.org "Quid quid Latine dictum sit, altum viditur" "Be excellent to each other ..." - Bill & Ted (Wyld Stallyns)

no unusual logs on maillog and dmesg also. Im thinking if its a connectivity problem on dovecot ? i tried one of my clients. downloading 40 mails, on the 20th mail it stops there and timeout.

Oh, on the 20th email! Well that changes EVERYTHING!

We need information to be able to help you. Use an email client like thunderbird and watch the email transaction, or use ethereal and sniff/follow the email transaction, or increase the logging on the server. Right now, you've not shown us how your server is configured. We have no logs to consult, only your word that they're okay, and we have no client to watch to see how the client treats the server. in short... GIVE US INFORMATION! You need to help us help you. -- "They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety'' Benjamin Franklin 1775

spart cus spake the following on 3/29/2006 10:07 PM:

*/Scott Silva ssilva@sgvwater.com/* wrote:

spart cus spake the following on 3/28/2006 6:38 PM:
> Hi To All Gurus,
> Im encountering some problems on my mailserver. When my clients are
> downloading their emails, when they reach around 15 mails downloaded
> somehow the connection stops. Im using Centos4.2, Postfix and Dovecot.
> Any one experiece these ?
>
Try and increase the timeouts for the clients.
Are you using Outlook on the clients?
There are some workarounds for outlook in Dovecot.

Yes some of my clients uses outlook.

When i check rpm -qa|grep dovecot, it says dovecot-0.99.11-2.EL4.1, so it means it uses .99 version.But when i check the dovecot.conf , at the beginning of the file

## Dovecot 1.0 configuration file

with this on the file

# outlook-pop3-no-nuls #client_workarounds = client_workarounds = outlook-pop3-no-nuls should i use this one ? and what does this config actually do so that i could explain to them the erros. thanks.

# outlook-pop3-no-nuls: # Outlook and Outlook Express hang if mails contain NUL characters. # This setting replaces them with 0x80 character.

It isn't exactly .99 it is .99 with the famous RedHat backporting of fixes, so it is somewhere between .99 stable and whatever 1.0 beta that they ported from. If you look in the stock config file, it is very well commented. I haven't enabled any workarounds, and haven't had any complaints since the initial migration and duplicated messages, but you can experiment with outlook_idle and outlook-pop3-no-nuls. Make sure that the locking is the same as your MUA and delivery agent. In CentOS 4 it should be fcntl. In CentOS 3 I think it should be flock, but don't hold me to it.