Packages in CentOS

[hiroyosi]

Hi I am trying to install pakages for ispCPOmega in CentOS5.

I followed INSTALL(ispCP ω (OMEGA) HowTo Install for CentOS 5).
But some packages could not be installed.

I tried as follows.

# yum install `cat ./docs/CentOS/centos-packages`

Some packages couldn't be installed as follows.

No package clamav-data available.
No package clamav-lib available.
No package clamav-server available.
No package clamav-update available.
No package cyrus available.
No package perl-SNMP_Session available.
No package perl-Term-ReadPassword available.
No package sasl-gssapi available.

These packages must be installed to use ispCPOmega?

[motokochan]

hiroyosi Wrote:I followed INSTALL(ispCP ω (OMEGA) HowTo Install for CentOS 5).
But some packages could not be installed.

--snip--

These packages must be installed to use ispCPOmega?

I currently recommend using the Fedora install package list. You'll have fewer missing packages reported. Of those, the clamav ones are old names. I believe you just need to make sure that freshclam and clamd are installed beyond everything else.

You'll also need to install the two missing perl modules via CPAN.

Likewise, use the CentOS versions of the Courier packages from thatfleminggent.com or you will break Postfix. Just navigate from the main page of that site to get the links for CentOS editions.

[obey]

motokochan Wrote:Likewise, use the CentOS versions of the Courier packages from thatfleminggent.com or you will break Postfix. Just navigate from the main page of that site to get the links for CentOS editions.

hi, i just downloaded and updated the packages with the thatfleminggent packages. i still got some serious issues.
1.: i cant log into the webmailer error:

Code:

ERROR: Connection dropped by IMAP server.

2.: i cant send any mails (wether through console mail command nor through a client)

i think postfix is totally fucked up. do you have any suggestions to get this system running without having to install everything again on a debian system?

greets obey

[motokochan]

First, check the mail log. If there are reasons IMAP is dropping the connection, it will be logged.

Next, edit /etc/postfix/master.conf and set the chroot section to "n" instead of "-". You should do it for at least the smtp service, but I believe one other might also need to be disabled for transport to work. Also, don't forget to set up /etc/aliases and run "newaliases" or you'll have errors when the smtpd service starts.

Edit: You need to set both smtp services (inet and unix type) to non-chroot. This should fix the queue from becoming backed up.

[obey]

motokochan Wrote:First, check the mail log. If there are reasons IMAP is dropping the connection, it will be logged.

the logged error is:

Code:

Jul 15 08:56:32 server1 postfix/smtpd[17766]: fatal: open database /etc/postfix/aliases.db: No such file or directory
Jul 15 08:56:33 server1 postfix/master[26145]: warning: process /usr/libexec/postfix/smtpd pid 17766 exit status 1
Jul 15 08:56:33 server1 postfix/master[26145]: warning: /usr/libexec/postfix/smtpd: bad command startup -- throttling

motokochan Wrote:Next, edit /etc/postfix/master.conf and set the chroot section to "n" instead of "-". You should do it for at least the smtp service, but I believe one other might also need to be disabled for transport to work.

done

motokochan Wrote:Also, don't forget to set up /etc/aliases and run "newaliases" or you'll have errors when the smtpd service starts.

what exactly is to be done to "set up" /etc/aliases ?

motokochan Wrote:Edit: You need to set both smtp services (inet and unix type) to non-chroot. This should fix the queue from becoming backed up.

isn't this done through the master.cf?

EDIT:
i just got another error in the log:

Code:

Jul 15 09:07:57 server1 authdaemond: ldap_simple_bind_s failed: Can't contact LDAP server
Jul 15 09:08:07 server1 authdaemond: ldap_simple_bind_s failed: Can't contact LDAP server

[motokochan]

obey Wrote:the logged error is:

Code:

Jul 15 08:56:32 server1 postfix/smtpd[17766]: fatal: open database /etc/postfix/aliases.db: No such file or directory
Jul 15 08:56:33 server1 postfix/master[26145]: warning: process /usr/libexec/postfix/smtpd pid 17766 exit status 1
Jul 15 08:56:33 server1 postfix/master[26145]: warning: /usr/libexec/postfix/smtpd: bad command startup -- throttling

That means you need to run the "newaliases" command to build the aliases map.

obey Wrote:

motokochan Wrote:Also, don't forget to set up /etc/aliases and run "newaliases" or you'll have errors when the smtpd service starts.

what exactly is to be done to "set up" /etc/aliases ?

Set the root account to go to a different address. Maybe you have a monitoring address for checking on server output? (If not, you should). In best practices, the root account shouldn't get mail delivered directly. Once done, run "newaliases" to build the map Postfix uses.

obey Wrote:

motokochan Wrote:Edit: You need to set both smtp services (inet and unix type) to non-chroot. This should fix the queue from becoming backed up.

isn't this done through the master.cf?

Yes, that goes with the above info I gave. Although I said master.conf previously, it is master.cf. Those are the only services you should need to make the edit to, I confirmed on one of my machines today.

obey Wrote:EDIT:
i just got another error in the log:

Code:

Jul 15 09:07:57 server1 authdaemond: ldap_simple_bind_s failed: Can't contact LDAP server
Jul 15 09:08:07 server1 authdaemond: ldap_simple_bind_s failed: Can't contact LDAP server

Hm. Check that courier-imap has the right config files. If you installed those RPMs after installing ispCP Omega, you might have the wrong config.

[obey]

obey Wrote:EDIT:
i just got another error in the log:

Code:

Jul 15 09:07:57 server1 authdaemond: ldap_simple_bind_s failed: Can't contact LDAP server
Jul 15 09:08:07 server1 authdaemond: ldap_simple_bind_s failed: Can't contact LDAP server

Hm. Check that courier-imap has the right config files. If you installed those RPMs after installing ispCP Omega, you might have the wrong config.
[/quote]

yeah, thats what i did...
i compared the configs to a (nearly) running debian system, and saw that some files seem to be missing
debian-sys:

Code:

obey:/etc/postfix# ls -al
total 100
drwxr-xr-x  4 root root  4096 2008-07-09 13:50 .
drwxr-xr-x 89 root root  4096 2008-06-06 15:26 ..
-rw-r--r--  1 root root   373 2008-05-21 23:19 dynamicmaps.cf
drwxr-xr-x  2 root root  4096 2008-07-09 11:31 ispcp
-rw-r--r--  1 root root  4810 2008-07-09 13:50 main.cf
-rw-r--r--  1 root root  1255 2008-05-22 00:59 main.cf_2008-05-21
-rw-r--r--  1 root root  5532 2008-05-24 01:51 master.cf
-rw-r--r--  1 root root   123 2008-07-09 09:35 mysql-virtual_gid_maps.cf
-rw-r--r--  1 root root   123 2008-07-09 09:36 mysql-virtual_uid_maps.cf
-rw-r--r--  1 root root 17975 2007-03-21 12:17 postfix-files
-rwxr-xr-x  1 root root  6840 2007-03-21 12:17 postfix-script
-rwxr-xr-x  1 root root 22239 2007-03-21 12:17 post-install
drwxr-xr-x  2 root root  4096 2008-07-09 09:40 sasl

centOS-sys:

Code:

[root@server1 postfix]# ls -al
insgesamt 316
drwxr-xr-x  3 root root  4096 15. Jul 09:17 .
drwxr-xr-x 82 root root  4096 15. Jul 09:18 ..
-rw-r--r--  1 root root 17832 21. Jan 2007  access
-rw-r--r--  1 root root  3550 21. Jan 2007  bounce.cf.default
-rw-r--r--  1 root root 11175 21. Jan 2007  canonical
-rw-r--r--  1 root root  9920 21. Jan 2007  generic
-rw-r--r--  1 root root 16838 21. Jan 2007  header_checks
drwxr-xr-x  2 root root  4096 29. Mai 14:24 ispcp
-rw-r--r--  1 root root 11942 21. Jan 2007  LICENSE
-rw-r--r--  1 root root  4848 15. Jul 09:17 main.cf
-rw-r--r--  1 root root 17981 21. Jan 2007  main.cf.default
-rw-r--r--  1 root root 24214  9. Jul 10:57 main.cf.rev
-rw-r--r--  1 root root   958 21. Jan 2007  makedefs.out
-rw-r--r--  1 root root  5531 15. Jul 09:07 master.cf
-rw-r--r--  1 root root 17639 21. Jan 2007  postfix-files
-rwxr-xr-x  1 root root  6366 21. Jan 2007  postfix-script
-rwxr-xr-x  1 root root 22564 21. Jan 2007  post-install
-rw-r--r--  1 root root  6805 21. Jan 2007  relocated
-rw-r--r--  1 root root  1629 21. Jan 2007  TLS_LICENSE
-rw-r--r--  1 root root 12081 21. Jan 2007  transport
-rw-r--r--  1 root root 11984 21. Jan 2007  virtual

and thx for your patience

[obey]

ok, i'm getting nearer to the solution...

where can i change the path courier looks for the userdb?

Code:

Jul 15 10:43:56 server1 authdaemond: Installing libauthuserdb
Jul 15 10:43:56 server1 authdaemond: Installation complete: authuserdb
Jul 15 10:44:11 server1 imapd: Connection, ip=[::ffff:127.0.0.1]
Jul 15 10:44:11 server1 authdaemond: received auth request, service=imap, authtype=login
Jul 15 10:44:11 server1 authdaemond: authuserdb: trying this module
Jul 15 10:44:11 server1 authdaemond: userdb: failed to open /etc/authlib/userdb.dat
Jul 15 10:44:11 server1 authdaemond: userdb: looking up 'test@test.domain.tld'
Jul 15 10:44:11 server1 imapd: LOGIN FAILED, user=test@test.domain.tld, ip=[::ffff:127.0.0.1]
Jul 15 10:44:11 server1 authdaemond: authuserdb: REJECT - try next module
Jul 15 10:44:11 server1 authdaemond: FAIL, all modules rejected
Jul 15 10:44:16 server1 imapd: LOGOUT, ip=[::ffff:127.0.0.1], rcvd=60, sent=332

OR

what values do i have to change to get this running in authmysqlrc?

Code:

Jul 15 10:36:22 server1 authdaemond: authmysqllib: connected. Versions: header 50022, client 50045, server 50045
Jul 15 10:36:22 server1 authdaemond: SQL query: SELECT username, password_enc, "", uid, gid, homedir, maildir, "", "", "" FROM mail_users WHERE username = 'test@test.domain.tld'
Jul 15 10:36:22 server1 authdaemond: mysql_query failed, reconnecting: Unknown column 'username' in 'field list'
Jul 15 10:36:22 server1 authdaemond: authmysqllib: connected. Versions: header 50022, client 50045, server 50045
Jul 15 10:36:22 server1 authdaemond: mysql_query failed second time, giving up: Unknown column 'username' in 'field list'
Jul 15 10:36:22 server1 authdaemond: authmysql: TEMPFAIL - no more modules will be tried
Jul 15 10:43:56 server1 authdaemond: stopping authdaemond children

[obey]

ok, i managed to get the authlib running. it was a linefeed (!?) in the userdb.dat.

i deleted it and ran makeuserdb again... and voila: it worked.

I am now able to log into the webmail service, but i still have got problems with delivering mails. Now the log tells me:

Code:

Jul 15 12:29:06 server1 courierd: started,id=042B0002.487C7BF2.0000526F,from=<root@server1.domain.tld>,module=esmtp,host=test.domain.tld,addr=<test@test.domain.tld>
Jul 15 12:29:06 server1 courieresmtp: id=042B0002.487C7BF2.0000526F,from=<root@server1.domain.tld>,addr=<test@test.domain.tld>: configuration error: mail loops back to myself (MX problem).
Jul 15 12:29:06 server1 courieresmtp: id=042B0002.487C7BF2.0000526F,from=<root@server1.domain.tld>,addr=<test@test.domain.tld>,status: failure

any idea whats wrong now?

[obey]

after a few hours of testing....

i fixed the loop error with a simple entry in the userdb. it seems postfix and courier aren't looking for the whole user entry (for example blubb@domain.com). they only look for blubb.
i inserted a single line with only the alias and without domain and the mail was delivered. what issue is that now?

UPDATE:
i fixed that one... it seems that the policyd-weight isn't running, so the process can't connect to port 12525. i just installed it from policyd-weight.org and postfix is up and running