After domain change, MTA fails to start: relayhost/myhostname identical

[randallz]

Starting zmmtaconfig...zmmtaconfig is already running.
postalias: fatal: myhostname and relayhost parameter settings must not be identical: mail.domain.org
postfix failed to start

The original domain name is in the process of being deprecated. I was able to add new accounts using the new domain, however I had never been able to get the logger accessible within the zimbraAdmin page, so I was taking this opportunity to attempt a fix the server status monitoring.

OLD: domain.com
NEW: example.com

The server had been mail.domain.com with an MX listing to get the domain.com email addresses sent correctly. 'www .domain .com' was hosted on a different server.

I changed the server name to example.com while telling it to skipusers

/opt/zimbra/libexec/zmsetservername --skipusers -o domain.com -n mail.example.com

Believing that the logger was not being recognized because I had not used the FQDN.

Anyway, that did not solve the logger problem, and neither did it when I switched with:

/opt/zimbra/libexec/zmsetservername --skipusers -n example.com

I decided to put off fixing the logger until later, since it is not immediately important, as I can check the status from the terminal.

I also needed to create a new certificate so, and wanted the error in mail and web clients to say that it was absent a trusted authority but that at least it was the address matched the certificate.

Deciding whether to use *.example.com or mail.example.com meant that I did this a couple times, and restarted the server each time.

During this process however, I did something that seemed to ruin the LDAP server. Amazingly through reinstalling I eventually seemed to restore LDAP to a working condition. Though somewhere in the process I also did:

/opt/zimbra/libexec/zmsetservername -n --force mail.example.org

However, now while it says that LDAP is working, postfix/MTA will not start, and claims that I have set a relayhost that is identical to myhostname. I do not believe that I ever had a relayhost. I am running CentOS 5.4 on a Linode VPS that could have automatically determined some network configuration including a relayhost through their kernel additions, but I highly doubt it.

So, I am stuck. I have tried removing the entry for a relayhost to blank in the main.cf file, however the value repopulates each time I restart.

I am not sure where I caused the problem and whether it is something that is still wrong with LDAP, a bad certificates authority configuration, or something that can be fixed within postfix config files.

Of note, I can not log in to the user web portal, it freezes after getting to the loading page. The zimbraAdmin page is partially accessible though I get this message now:

JavaScript error encountered in method ZaZimbraAdmin.prototype._getLoginMsgPanel message: ZaApp.getInstance().getCurrentController() is undefined fileName: https:// mail. example.org:7071/zimbraAdmin/js/Admin_all.js.zgz?v=100820055815 lineNumber: 21838 stack: ("name","example.org")@https://mail.example.org:7071/zimbraAdmin/js/Admin_all.js.zgz?v=100820055815:21838 ("...

That I can log in to the zimbraAdmin page suggests that the LDAP is not completely broken. However, if I try to "View Mail" for one of the users' accounts from the admin page, it looks like th problem is that it is still using the old domain. The error is:

Server error encountered Message: operation sent to wrong host (you want 'domain.org') Error code: service.WRONG_HOST Details:soap:Sender

---- ---- ----

So, I have put a lot of hours into trying to fix whatever I broke, and I have done it without a careful plan. Hopefully the symptoms sound like something obvious to someone, but at this point I am almost at a dead end, since I am not all that sure how to narrow down the problem to LDAP ,certificates, or postfix.

Thanks,
Randall

[borngunners]

Hi...
I am having the same issue like yours... What did you do to resolve it?

[phoenix]

Originally Posted by borngunners

I am having the same issue like yours... What did you do to resolve it?

What exactly have you done to your server? What errors are you seeing in the log files?

[borngunners]

I changed the domain name from zimbra.health.umd.edu to health.umd.edu. In the server status, I am seeing both servers in there and the "zimbra.health.umd.edu" have staled data as indicated in the status. Example:
maillog November 27 (data is stale)
.........
.........
........
.......
spell November 27 (data is stale)

[phoenix]

Have you checked to see if the public service hostname is set correctly?

[borngunners]

Thanks for your response... I found out that there was no public service hostname setup. I have setup the public service hostname and restart zimbra
using the following command

zmcontrol restart

Still no luck

[phoenix]

Originally Posted by borngunners

Thanks for your response... I found out that there was no public service hostname setup. I have setup the public service hostname and restart zimbra
using the following command

What have you set it to and is that hostname able to be resolved via DNS? Check it using the commands in the 'Verify...' section of the Split DNS article, post the output here if you need to.

[borngunners]

Below is the output of the split dns check:

host `hostname`

health.umd.edu has address 128.8.222.205
health.umd.edu mail is handled by 5 health.umd.edu.
health.umd.edu mail is handled by 15 mail.health.umd.edu.

ee /etc/resolv.conf

search umd.edu
nameserver 128.8.5.2
nameserver 128.8.74.2
nameserver 128.8.76.2

dig health.umd.edu mx

;; global options: printcmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 25291
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 2, AUTHORITY: 3, ADDITIONAL: 5

;; QUESTION SECTION:
;health.umd.edu. IN MX

;; ANSWER SECTION:
health.umd.edu. 300 IN MX 5 health.umd.edu.
health.umd.edu. 300 IN MX 15 mail.health.umd.edu.

;; AUTHORITY SECTION:
umd.edu. 60000 IN NS ns2.umd.edu.
umd.edu. 60000 IN NS noc.umd.edu.
umd.edu. 60000 IN NS ns1.umd.edu.

;; ADDITIONAL SECTION:
health.umd.edu. 300 IN A 128.8.222.205
mail.health.umd.edu. 300 IN A 128.8.222.203
noc.umd.edu. 60000 IN A 128.8.5.2
ns1.umd.edu. 60000 IN A 128.8.74.2
ns2.umd.edu. 60000 IN A 128.8.76.2

;; Query time: 2 msec
;; SERVER: 128.8.5.2#53(128.8.5.2)
;; WHEN: Sun Nov 28 07:19:10 2010
;; MSG SIZE rcvd: 203

dig health.umd.edu any

; <<>> DiG 9.4.2-P2 <<>> health.umd.edu any
;; global options: printcmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 3330
;; flags: qr aa rd ra; QUERY: 1, ANSWER: 3, AUTHORITY: 3, ADDITIONAL: 4

;; QUESTION SECTION:
;health.umd.edu. IN ANY

;; ANSWER SECTION:
health.umd.edu. 300 IN A 128.8.222.205
health.umd.edu. 300 IN MX 5 health.umd.edu.
health.umd.edu. 300 IN MX 15 mail.health.umd.edu.

;; AUTHORITY SECTION:
umd.edu. 60000 IN NS noc.umd.edu.
umd.edu. 60000 IN NS ns2.umd.edu.
umd.edu. 60000 IN NS ns1.umd.edu.

;; ADDITIONAL SECTION:
mail.health.umd.edu. 300 IN A 128.8.222.203
noc.umd.edu. 60000 IN A 128.8.5.2
ns1.umd.edu. 60000 IN A 128.8.74.2
ns2.umd.edu. 60000 IN A 128.8.76.2

;; Query time: 1 msec
;; SERVER: 128.8.5.2#53(128.8.5.2)
;; WHEN: Sun Nov 28 07:22:08 2010
;; MSG SIZE rcvd: 203

[phoenix]

What's in your hosts file?

[borngunners]

127.0.0.1 localhost.localdomain localhost
128.8.222.205 health.umd.edu health

# The following lines are desirable for IPv6 capable hosts
::1 ip6-localhost ip6-loopback
fe00::0 ip6-localnet
ff00::0 ip6-mcastprefix
ff02::1 ip6-allnodes
ff02::2 ip6-allrouters
ff02::3 ip6-allhosts