Starting LDAP - FAIL. Ubuntu, dapper

[cshepherd]

Hey guys,

I have been trying for days to get Zimbra 3.1.3 installed on Dapper (I have previously setup 3.1.0 on Breezy a couple times using the coerce.sh script).

Also of note, I'm using VServers as well.

These are the steps I've been following;

Installed cron curl fetchmail file ftp libpcre3 libgmp3-dev libgmp3c2 libxml2 lynx make ncftp patch perl psmisc ssh unzip wget zip.

I then followed this guide for creating an Ubuntu package.

Setting Defaults... takes around 15-20mins to time out before it comes up with the next step of changing the domain (All A and MX records resolve correctly). I don't know if this is a separate problem or not...

Searching the forums I thought it may have been sudo, but this wasn't the case. I tried starting LDAP manually and it just freezes.

The only logs I can find are in the Zimbra setup logs;

/opt/zimbra/libexec/zmldapinit: line 61: /opt/zimbra/conf/slapd.conf.in: Permission denied
/opt/zimbra/lib/conf/zimbra-hsm.schema: line 1: ObjectIdentifier "ZimbraHsmAttrType" previously defined "1.3.6.1.4.1.19348.2.4"ERROR - failed to start slapd

This happens on a clean install or a retried install.

Any ideas? Am I supplying the correct info?

Any help would be very much appreciated.

Thanks,

Craig

[KevinH]

Are you trying to install a Network version or OSS? That HSM attr is part of the Network Edition so maybe you've got a mixed environment and not a clean OSS or Network setup.

[cshepherd]

Definitely OSS as I don't remember going through any registration process that it seems you need for the network version.

Here is the file name zcs-3.1.3_GA_474.DEBIAN3.1.tgz

[cshepherd]

Ideas anyone?

[phoenix]

I'm afraid my only suggestion would be to remove zimbra with './install.sh -u', kill any running processes and then do a re-install and see what happens. Have you already tried that?

[cshepherd]

Being that it's running in a VServer it's no effort for me to wipe out the entire server and start from scratch. I have started from scratch multiple times.

If it helps, I have also built a Debian 'Sarge' VServer and I'm having the same issues.

[cshepherd]

I have replicated this problem over and over again. Has anyone else experienced this. I have tried brand new vservers of Debian Sarge and Ubuntu Dapper with 3.1.3 Debian build of Zimbra

I've found something extra in /var/log/syslog that may help. What on earth is liquidsys.com?

Jun 29 11:40:26 mail slapd[18261]: @(#) $OpenLDAP: slapd 2.3.21 (Jun  1 2006 18:12:55) $ ^Iroot@build-debian.liquidsys.com:/home/build/p4/CRAY/ThirdParty/openldap/openldap-2.3.21/servers/slapd
Jun 29 11:40:26 mail slapd[18261]: daemon: bind(7) failed errno=99 (Cannot assign requested address)
Jun 29 11:40:26 mail slapd[18261]: slapd stopped.

[phoenix]

Is your hosts file set correctly and the server resolvable by DNS? IIRC, liquidsys.com was the name Zimbra used before they became Zimbra.

[cshepherd]

I have an internal DNS server setup that is resolving correctly and even all the PTRs are setup and resolving. The MX is also coming back correctly when I check with dig.

I've given up with 3.1.3 and gone back to 3.1.0 which, while it paused for up to an hour on some steps, it installed successfully.

[KevinH]

liquidsys.com was our stealth name/domain before we launched. We still use it internally for some of the servers.

The bind error looks like there is another process already bound to one of the LDAP ports. Can you verify that no other LDAP service is running.