1 (edited by Rashef 2009-05-22 23:04:37)

Topic: Slow/refused connections and other problems.

I've just installed iRedMail package on a fresh CentOS 5.3.
The server is connected into the DMZ (192.168.10.x). Clients are external IPs and users from LAN (192.168.0.x). Most of users are members of an ADS domain. Some users (Mac and Linux users) are out of the domain (same subnet).

My problems started when I attempted to migrate the mailboxes. I used a little tool (IMAPcopy) but it fails. So I'm trying to move messages by hand. I configured both old (hMailServer) and new (iRedMail suite) IMAP servers and started moving folders or mails.

Sometimes the moving starts but randomly hangs. Sometimes it doesn't start (Error: no output stream). I used Evolution from a Linux box.

Moreover, some users can connect to roundcube webmail very quickly: pages are vary fast (loaded in less then a second); other users have got to wait ten or more seconds to have the page loaded. I cannot recognize any link between users w/ problems and domain.

This problem occurred with:
some ADS users
a linux box

The problem didn't occurr with:
other ADS users
a Mac device

What to check?

Thanks in advance!

P.S.: I hope I will able to solve these issues... You made a great job guys! smile

EDIT
Configured with MySQL....

----

Spider Email Archiver: On-Premises, lightweight email archiving software developed by iRedMail team. Supports Amazon S3 compatible storage and custom branding.

2

Re: Slow/refused connections and other problems.

You didn't post any log message here.

Suggestions:

  • Enable IMAP server debug mode in /etc/dovecot.conf (mail_debug = yes), resetart dovecot server and monitor /var/log/dovecot.log:

  • Monitory MySQL queries from postfix and dovecot.

And don't forget to monitory client application log.

3

Re: Slow/refused connections and other problems.

Thanks!
I'm monitoring those logfiles. In the meanwhile I noticed that when the machine hangs it seems like it is no more available in the network. When it hangs this appears into /var/log/messages:

May 22 18:05:37 h100 kernel: eth0: Detected Tx Unit Hang:
May 22 18:05:37 h100 kernel:   TDH                  <4b>
May 22 18:05:37 h100 kernel:   TDT                  <4f>
May 22 18:05:37 h100 kernel:   next_to_use          <4f>
May 22 18:05:37 h100 kernel:   next_to_clean        <49>
May 22 18:05:37 h100 kernel: buffer_info[next_to_clean]:
May 22 18:05:37 h100 kernel:   time_stamp           <6750d8b>
May 22 18:05:37 h100 kernel:   next_to_watch        <4b>
May 22 18:05:37 h100 kernel:   jiffies              <6751768>
May 22 18:05:37 h100 kernel:   next_to_watch.status <0>
May 22 18:05:39 h100 kernel: eth0: Detected Tx Unit Hang:
May 22 18:05:39 h100 kernel:   TDH                  <4b>
May 22 18:05:39 h100 kernel:   TDT                  <4f>
May 22 18:05:39 h100 kernel:   next_to_use          <4f>
May 22 18:05:39 h100 kernel:   next_to_clean        <49>
May 22 18:05:39 h100 kernel: buffer_info[next_to_clean]:
May 22 18:05:39 h100 kernel:   time_stamp           <6750d8b>
May 22 18:05:39 h100 kernel:   next_to_watch        <4b>
May 22 18:05:39 h100 kernel:   jiffies              <6751f39>
May 22 18:05:39 h100 kernel:   next_to_watch.status <0>
May 22 18:05:41 h100 kernel: eth0: Detected Tx Unit Hang:
May 22 18:05:41 h100 kernel:   TDH                  <4b>
May 22 18:05:41 h100 kernel:   TDT                  <4f>
May 22 18:05:41 h100 kernel:   next_to_use          <4f>
May 22 18:05:41 h100 kernel:   next_to_clean        <49>
May 22 18:05:41 h100 kernel: buffer_info[next_to_clean]:
May 22 18:05:41 h100 kernel:   time_stamp           <6750d8b>
May 22 18:05:41 h100 kernel:   next_to_watch        <4b>
May 22 18:05:41 h100 kernel:   jiffies              <6752708>
May 22 18:05:41 h100 kernel:   next_to_watch.status <0>
May 22 18:05:43 h100 kernel: eth0: Detected Tx Unit Hang:
May 22 18:05:43 h100 kernel:   TDH                  <4b>
May 22 18:05:43 h100 kernel:   TDT                  <4f>
May 22 18:05:43 h100 kernel:   next_to_use          <4f>
May 22 18:05:43 h100 kernel:   next_to_clean        <49>
May 22 18:05:43 h100 kernel: buffer_info[next_to_clean]:
May 22 18:05:43 h100 kernel:   time_stamp           <6750d8b>
May 22 18:05:43 h100 kernel:   next_to_watch        <4b>
May 22 18:05:43 h100 kernel:   jiffies              <6752ed8>
May 22 18:05:43 h100 kernel:   next_to_watch.status <0>
May 22 18:05:43 h100 kernel: NETDEV WATCHDOG: eth0: transmit timed out

4

Re: Slow/refused connections and other problems.

A network issue? Or hardware/kernel driver issue?

5

Re: Slow/refused connections and other problems.

ZhangHuangbin wrote:

A network issue? Or hardware/kernel driver issue?

I dunno... I'm using two card:

0d:00.0 Ethernet controller: Intel Corporation 82573E Gigabit Ethernet Controller (Copper) (rev 03)
0e:00.0 Ethernet controller: Intel Corporation 82573L Gigabit Ethernet Controller

the first one is up and linked to the switch... I will try the second one... :-/

6 (edited by Rashef 2009-05-23 00:49:17)

Re: Slow/refused connections and other problems.

I cannot understand what happens but the server is dramatically slow even with local login... It has a 4 GB RAM, a Intel Dual Core 3 GHz and 750 GB SATA RAID 5... It should not be a resource matter...

Now I have half of my mailboxes on this server, and the other half on the old one... Good luck to me... sad

7

Re: Slow/refused connections and other problems.

Ok I solved.
The issue was related to that damn Intel Pro gigabit network card and its buggy e1000 kernel module... :-|