| *SOLVED* Q: Postfix Kills Interface of VirtualHost .. Install Hints? [message #5658] |
Fri, 25 August 2006 15:57  |
tchipman
Messages: 28
Registered: June 2006
|
Junior Member |
|
|
I wanted to ask if anyone has hints about how to install Postfix such that grief doesn't ensue.
Currently, my config / issue is thus:
-CentOS 4.3 32-bit Hardware host platform
-Latest OpenVZ Stable kernel installed, running
-CentOS "Default" 32-bit (or "minimal") Virtual Host Environment
If I create a clean virtual machine, then install postfix and use "system-switch-mail" to disable sendmail / enable postfix..
Postfix seems to be having trouble because in the config is an option which specifies what interface to listen on. By default it is "localhost". I've tried changing it to
lo
venet0
venet0:0
and also tried adding entries to /etc/hosts for the above .. to reduce complaint from postfix when I do a "service postfix stop ; service postfix start" in the VZHost.
Alas, I get interesting results:
(1) Postfix is pissed off
(2) My interface falls off the map, ie:
BEFORE postfix is running:
---paste---
root@nema log]# ifconfig -a
lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:28 errors:0 dropped:0 overruns:0 frame:0
TX packets:28 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:3344 (3.2 KiB) TX bytes:3344 (3.2 KiB)
venet0 Link encap:UNSPEC HWaddr 00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00
inet addr:127.0.0.1 P-t-P:127.0.0.1 Bcast:0.0.0.0 Mask:255.255.255.255
UP BROADCAST POINTOPOINT RUNNING NOARP MTU:1500 Metric:1
RX packets:488 errors:0 dropped:0 overruns:0 frame:0
TX packets:287 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:36797 (35.9 KiB) TX bytes:41549 (40.5 KiB)
venet0:0 Link encap:UNSPEC HWaddr 00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00
inet addr:1X9.1X3.X3.2X9 P-t-P:1X9.1X3.X3.2X9 Bcast:1X9.1X3.X3.2X9 Mask:255.255.255.255
UP BROADCAST POINTOPOINT RUNNING NOARP MTU:1500 Metric:1
--end-paste--
(note "X" is added to make my IP private here in the posting, it is a regular number "for real"...
AFTER postfix is running:
--paste---
[root@hydra cache]# vzctl enter 239
entered into VPS 239
-bash-3.00# ifconfig -a
lo Link encap:Local Loopback
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:28 errors:0 dropped:0 overruns:0 frame:0
TX packets:28 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:3344 (3.2 KiB) TX bytes:3344 (3.2 KiB)
venet0 Link encap:UNSPEC HWaddr 00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00
UP BROADCAST POINTOPOINT RUNNING NOARP MTU:1500 Metric:1
RX packets:4 errors:0 dropped:0 overruns:0 frame:0
TX packets:3 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:216 (216.0 b) TX bytes:164 (164.0 b)
venet0:0 Link encap:UNSPEC HWaddr 00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00
UP BROADCAST POINTOPOINT RUNNING NOARP MTU:1500 Metric:1
-bash-3.00#
ie, no IP associated for venet0:0.
Ugh!
It strikes me as particularly interesting that postfix not only chokes, but kills the interface as well.
If anyone has comments on this, it would be greatly appreciated.
Many thanks,
--Tim
[Updated on: Fri, 25 August 2006 19:21] by Moderator Report message to a moderator |
|
|
|
| Re: Q: Postfix Kills Interface of VirtualHost .. Install Hints? [message #5660 is a reply to message #5658] |
Fri, 25 August 2006 16:15  |
tchipman
Messages: 28
Registered: June 2006
|
Junior Member |
|
|
DOH. Sorry folks, I forgot the magic rule: RTFM, and after that, google search your keywords. THEN, and only then, may you post innane questions to the forum.
Solution (thanks, google) - was hinted in the forum elsewhere.
-start the VZHost
-take note of problems clearly cropping up in /proc/user_beancounters as follows:
PROBLEM LAYER ONE:
420 cat /proc/user_beancounters
numothersock 4 80 80 80 81
So, increased this parameter as follwos:
419 vzctl 239 --numothersock 120 --save
-stopped and started my Vhost. Check Beancounters.
PROBLEM LAYER TWO:
426 cat /proc/user_beancounters
READS
othersockbuf 128788 133452 132096 336896 123
So, increased this parameter as follows:
428 vzctl set 239 --othersockbuf 512000 --save
-stop and start the Vhost. Check Beancounters. NO complaints obvious.
Enter the VZhost. Postfix is running. Send a test message. It gets out, postfix is happy.
Subsequently: I tweaked the postfix config such that my interfac e was specified as "localhost" (my preferred config, nobody outside of this virtual-box should be relaying mail or sending mail in for that matter)
This also works fine.
So, I'm happy.
Many thanks all 
---Tim
|
|
|
|
| Re: *SOLVED* Q: Postfix Kills Interface of VirtualHost .. Install Hints? [message #6708 is a reply to message #5658] |
Thu, 21 September 2006 15:05 |
hanksteronimous
Messages: 1
Registered: September 2006
|
Junior Member |
|
|
Having the same problem and have used this post to solve the majority of issues, however, after doing the above I find the following:
from the HM and NOT looged in as root (or Su'ed)
nmap VPSName finds port 25 is open and
telnet VPSName 25 gets a response.
from the HM and logged in as root (or SU'ed)
nmap complains that "the Note: Host seems down. If it is really up, but blocking our ping probes, try -P0"
telnet VPSName gets a response...
from anywhere else on the network
nmap VPSName complains (see above)
telnet VPSName 25 gets "could not open connection to the host on port 25: connect failed"
yet ping VPSName does get a response
Have verified that this IP address is NOT in use elsewhere on the network.
Any ideas???
|
|
|
|
OpenVZ Forum
Members
Search
Help
Register
Login
Home
