IMAP, Apache, some ping Fail on New Hardware??

Installing, Configuring, Troubleshooting server daemons such as Web and Mail
dunwell
Posts: 116
Joined: 2010/12/20 18:49:52
Location: Colorado
Contact:

IMAP, Apache, some ping Fail on New Hardware??

Post by dunwell » 2014/07/23 23:13:25

I'm running COS5.Final I recently moved my working RAID1 disks to a new Lenovo server box. There were a few glitches related to the new hardware that were expected, you can see a rundown on that at
viewtopic.php?f=19&t=47315&p=201568#p201568
and the last post for a summary of steps.

However I've run into some problems that I didn't think would be hardware related to new drivers, new MAC, etc. All kind of relate to the ethernet connect though.

I am not a guru but a struggling sysadmin for my home server. So please err on the side of more detail and verbosity with any response. TIA

- IMAP and POP3 don't seem to work. When I look at the config files for dovecot (dovecot.conf and xinetd.d/imap) I don't see anything that I should change. If I run the GUI thing under System Admin, Server Settings, Services I see that dovecot shows as Stopped. If I click on the restart it says it re-started OK but still shows as Stopped. If I telnet to the pop3 port I see

Code: Select all

> telnet localhost pop3
Trying 127.0.0.1...
Connected to localhost.localdomain (127.0.0.1).
Escape character is '^]'.
+OK POP3 localhost.localdomain 2006k.101 server ready
user BlatherHere
-ERR Unknown AUTHORIZATION state command
so it really isn't talking. The usual things like a restart of a full reboot of the machine make no change. Along with the IMPA/POP3 is that ssh doesn't work coming in. You can ssh out OK but you can not ssh in.

Related is the Apache stuff. I can not connect to any of my web pages. Again, searching through the config files doesn't bring up anything that looks like it should be changed.

Possibly related is pings. I can ping out of course and use browsers out. If I ping "MyServerNameHere.com" from my laptop which is on the inside net on my router as 192.168.1.25, it will connect to the server. So going out through the router to a DNS and back in works. If however I ping to 192.168.1.21 (the server inside IP) then it fails. I can SAMBA connect to the server from the laptop OK though.

So something or maybe some things is/are not quite right and I don't know where else to look. They are all sort of ethernet connect related and they all happened with the hardware move. All were working fine on the old box.

Any thoughts appreciated. If this should be over on the Networking forum, Admin please move.
Alan D.

User avatar
TrevorH
Forum Moderator
Posts: 26887
Joined: 2009/09/24 10:40:56
Location: Brighton, UK

Re: IMAP, Apache, some ping Fail on New Hardware??

Post by TrevorH » 2014/07/23 23:44:42

Post the output of iptables-save and also lspci -nn | grep -i net

Did your server pick up a different IP address to the old one using DHCP? DHCP hands out IP addresses based on the MAC address of the ethernet card so changing cards will cause it to pick up a new ip address.
CentOS 5 died in March 2017 - migrate NOW!
CentOS 6 goes EOL sooner rather than later, get upgrading!
Full time Geek, part time moderator. Use the FAQ Luke

dunwell
Posts: 116
Joined: 2010/12/20 18:49:52
Location: Colorado
Contact:

Re: IMAP, Apache, some ping Fail on New Hardware??

Post by dunwell » 2014/07/24 00:23:14

Howdy again Trevor,

output of iptables-save attached. output from lspci was nothing at all. I did a redirect to a file and it is 0 in size.

The fixed IP for my web pages is actually outlooking side of my Linksys router. It uses the pass-through DMZ to the server's internal IP address of 192.168.1.21 so none of that should have changed. It should still be using the MAC of the Linsys.
Attachments
iptables-save.txt
output from iptables-save
(3.42 KiB) Downloaded 151 times

User avatar
TrevorH
Forum Moderator
Posts: 26887
Joined: 2009/09/24 10:40:56
Location: Brighton, UK

Re: IMAP, Apache, some ping Fail on New Hardware??

Post by TrevorH » 2014/07/24 00:30:46

Is the pciutils package installed? That's what provides the lspci command.
CentOS 5 died in March 2017 - migrate NOW!
CentOS 6 goes EOL sooner rather than later, get upgrading!
Full time Geek, part time moderator. Use the FAQ Luke

dunwell
Posts: 116
Joined: 2010/12/20 18:49:52
Location: Colorado
Contact:

Re: IMAP, Apache, some ping Fail on New Hardware??

Post by dunwell » 2014/07/24 01:40:55

TrevorH wrote:Is the pciutils package installed? That's what provides the lspci command.
Yes rpm reports V 3.1.7-5.el5

User avatar
TrevorH
Forum Moderator
Posts: 26887
Joined: 2009/09/24 10:40:56
Location: Brighton, UK

Re: IMAP, Apache, some ping Fail on New Hardware??

Post by TrevorH » 2014/07/24 09:46:57

You could try lspci -nn | grep -i eth and see if that finds it, otherwise run lspci -nn and manually inspect the devices it finds and see if you can see the cards that way.
CentOS 5 died in March 2017 - migrate NOW!
CentOS 6 goes EOL sooner rather than later, get upgrading!
Full time Geek, part time moderator. Use the FAQ Luke

dunwell
Posts: 116
Joined: 2010/12/20 18:49:52
Location: Colorado
Contact:

Re: IMAP, Apache, some ping Fail on New Hardware??

Post by dunwell » 2014/07/24 14:53:32

Dyslexics of the world untie! <sigh> I looked back and I was typing -nm rather than nn.

Output from the real command

Code: Select all

   lspci -nn | grep -i net
is attached.
Attachments
lspci.txt
(103 Bytes) Downloaded 194 times

User avatar
TrevorH
Forum Moderator
Posts: 26887
Joined: 2009/09/24 10:40:56
Location: Brighton, UK

Re: IMAP, Apache, some ping Fail on New Hardware??

Post by TrevorH » 2014/07/24 16:23:26

So you have an Intel i217 network card 8086:153a which is supported by the e1000e module. Those cards are usually fairly reliable but perhaps it's worth trying the kmod-e1000e package from the ELRepo.

I don't really have a clue what's wrong to be honest.
CentOS 5 died in March 2017 - migrate NOW!
CentOS 6 goes EOL sooner rather than later, get upgrading!
Full time Geek, part time moderator. Use the FAQ Luke

dunwell
Posts: 116
Joined: 2010/12/20 18:49:52
Location: Colorado
Contact:

Re: IMAP, Apache, some ping Fail on New Hardware??

Post by dunwell » 2014/07/24 16:58:37

There actually is no card, it is all off the motherboard. I'm tracking a bit about the internal IP address for the eth0. It looks like it has been assigned an internal IP inet addr of 192.168.1.102 when I look with ifconfig. I think that is supposed to be my assignment of 192.168.1.21 for all the other stuff to find it properly. I'm diging in to try to find where I set/change that. The files down in

Code: Select all

   /etc/sysconfig/networking/.../ifcfg.eth0
in the various branches there all are very brief and don't contain any IP info. They all look like

Code: Select all

# Intel Corporation Ethernet Connection I217-LM
DEVICE=eth0
BOOTPROTO=dhcp
ONBOOT=yes
HWADDR=44:39:c4:39:97:8c
That is where I went to reset the MAC address to get the eth0 to work right after the changeover to the new hardware. So I'm not sure how I change the IP for eth0.

dunwell
Posts: 116
Joined: 2010/12/20 18:49:52
Location: Colorado
Contact:

Re: IMAP, Apache, some ping Fail on New Hardware??

Post by dunwell » 2014/07/24 17:58:29

OK, I was being a bit stupid. I went back into the ifcfg-eth0 file area in

Code: Select all

  \etc\sysconfig\networking\profiles\default
and looked at the old one I'd saved off as "-Dell", this was the settings I had on the old box. There were a lot more lines there including setting the IP. So I just copied that, modified the MAC address line to be the new MAC and the comment header line for the new hardware and now it _seems_ to be all working. The file now looks like

Code: Select all

# Intel Corporation Ethernet Connection I217-LM
DEVICE=eth0
BOOTPROTO=none
BROADCAST=192.168.1.255
HWADDR=44:39:C4:39:97:8C
IPADDR=192.168.1.21
NETMASK=255.255.255.0
NETWORK=192.168.1.0
ONBOOT=yes
GATEWAY=192.168.1.10
TYPE=Ethernet
USERCTL=no
IPV6INIT=no
PEERDNS=yes
After a reboot this all seems to work, including IMAP from my laptop to the server. However on reboot I saw a Fail line for Dovecot saying the same thing as from the command line

Code: Select all

   Failed - Address already in use
Further investigation shows that the fail is on port 143 and xinetd is grabbing it before dovecot. Again, not sure why or how and not sure how to correct this or even if I should. I do have the UWashington sendmail installed and my notes say it is taking precedence over dovecot but it also says that I need the dovecot config file. Further, I don't see how a hardware change would modify any of those settings.

Post Reply

Return to “CentOS 5 - Server Support”