Issue with Multi-Master LDAP Servers

[vnngoanhtuan]

Hi All,

Model: 02 master ldap server.
- Master ldap server 01: zimbra-ldap.mydomain.com
- Master ldap server 02: zimbra-ldap2.mydomain.com

Currently, I meet a issue when zimbra-ldap.mydomain.com was broken. After fixed the server zimbra-ldap.mydomain.com. I run command as below:

- On zimbra-ldap2.mydomain.com:
$/opt/zimbra/libexec/zmslapcat /tmp/Backup
$cd /tmp/Backup
$ls -l
-rw-r----- 1 zimbra zimbra 12234190 Feb 20 09:05 ldap.bak
-rw-r----- 1 zimbra zimbra 12234190 Feb 20 09:05 ldap.bak.20190220090549

copy ldap.bak.20190220090549 from zimbra-ldap2.mydomain.com to zimbra-ldap.mydomain.com

- On zimbra-ldap.mydomain.com:
$su zimbra
$zmcontrol stop
$rm -rf /opt/zimbra/data/ldap/config/*
$rm -rf /opt/zimbra/data/ldap/mdb/*
$mkdir -p /opt/zimbra/data/ldap/mdb/db /opt/zimbra/data/ldap/mdb/logs
$chown -R zimbra:zimbra /opt/zimbra/data/ldap

- Restore configuration file
$/opt/zimbra/libexec/zmslapadd -c /tmp/Backup/ldap-config.bak (Backup config on zimbra-ldap.mydomain.com)

- Restore LDAP database file
$/opt/zimbra/libexec/zmslapadd /tmp/Backup/ldap.bak.20190220090549

$zmcontrol start

---- ERROR -----
Host zimbra-ldap.mydomain.com
Starting ldap...Done.
Size error: Unable to determine enabled services from ldap.
Enabled services read from cache. Service list may be inaccurate.
Starting zmconfigd...Failed.
Starting zmconfigd.../opt/zimbra/bin/zmconfigdctl: line 82: kill: (2708) - No such process
failed.


Starting snmp...Done.
Starting stats...Done.

$cat zmconfigd-log4j.log

---- LOG -----
2019-02-20 03:42:11,449 FATAL [sc] [] system - Could not find an LDAP entry for server 'zimbra-ldap.mydomain.com'
2019-02-20 03:43:21,468 FATAL [sc] [] system - Could not find an LDAP entry for server 'zimbra-ldap.mydomain.com'
2019-02-20 08:13:29,128 FATAL [sc] [] system - Could not find an LDAP entry for server 'zimbra-ldap.mydomain.com'
2019-02-20 09:04:18,466 FATAL [sc] [] system - Could not find an LDAP entry for server 'zimbra-ldap.mydomain.com'
2019-02-20 10:57:02,190 FATAL [sc] [] system - Could not find an LDAP entry for server 'zimbra-ldap.mydomain.com'
2019-02-20 11:06:19,788 FATAL [sc] [] system - Could not find an LDAP entry for server 'zimbra-ldap.mydomain.com'
2019-02-20 11:23:10,319 FATAL [sc] [] system - Could not find an LDAP entry for server 'zimbra-ldap.mydomain.com'
2019-02-20 11:58:31,812 FATAL [sc] [] system - Could not find an LDAP entry for server 'zimbra-ldap.mydomain.com'
2019-02-20 13:44:38,591 FATAL [sc] [] system - Could not find an LDAP entry for server 'zimbra-ldap.mydomain.com'
2019-02-20 14:15:21,160 FATAL [sc] [] system - Could not find an LDAP entry for server 'zimbra-ldap.mydomain.com'
2019-02-20 15:15:19,550 FATAL [sc] [] system - Could not find an LDAP entry for server 'zimbra-ldap.mydomain.com'
2019-02-20 15:37:55,389 FATAL [sc] [] system - Could not find an LDAP entry for server 'zimbra-ldap.mydomain.com'
2019-02-20 16:13:19,003 FATAL [sc] [] system - Could not find an LDAP entry for server 'zimbra-ldap.mydomain.com'

$zmupdateauthkeys did not see zimbra-ldap.mydomain.com in list.

Grep ldap database and did not see the dn and cn : zimbra-ldap.mydomain.com in this database.

$ grep -r zimbra-ldap ldap.bak.20190220090549
dn: cn=zimbra-ldap2.mydomain.com,cn=servers,cn=zimbra
cn: zimbra-ldap2.mydomain.com
zimbraServiceHostname: zimbra-ldap2.mydomain.com
1k6................./Gj zimbra-ldap2.mydomain.com

Can you please give me more idea and solution to fix the issues?

Thanks,
Tuan

[pup_seba]

Hi mate,

I remember this one time MMR was broken in a production site (mmr was just a new zimbra thing during that time and it deployed broken out of the box), lucky for me Quanah connected remotly and dude...that guy knew what he was doing I don't remember the exact steps, but basically he removed the broken ldap database, and rebuild the replication between both servers.

So, maybe just rebuilding the db from replication instead of backup/restore is an option in your case too? If I were in your situation, I would try things like:
- try to manually rebuild the replication (removing db, and exploring zimbra commands to recreate a replication agreement).
- re-running zmsetup.pl in the broken node.

Sorry I don't have the expreience to give you a "step by step" recepie, but I imagine doing such things could be easier than a backup/restore operation.

Hope you manage to fix the issue!

[vnngoanhtuan]

Hi Pup,

Thanks for your informations. I was fixed the issues as I follow the installation guideline to delete zimbra-ldap.mydomain out of mmr and re-install as new server.

Thanks,
Tuan Ngo