- 1 Scenario I - Single Server
- 2 Scenario II - Mysql Server with corrupted mysql data that is beyond repair
- 3 Scenario III - Multi Server: Restore LDAP Service
- 3.1 Method A
- 3.2 Method B
- 3.3 If ldap is not starting, use the long command to start ldap
- 3.4 Section III: Summary
Scenario I - Single Server
You have been running ZCS-88 using the Next Generation Suite, and you have a dedicated disk for /opt/zimbra/, something like...
/dev/mapper/vg_zcs-lv_root xxG xxxG xxG xx% / tmpfs /dev/shm /dev/sda1 /boot /dev/sda2 /opt/zimbra/
But something has happened to the server node, effecting the operating system, which will require a reinstall of the operating system, and ZCS-88.
Reinstall the Operating System
When you reinstall the operating system, make sure the disk containing /opt/zimbra/ is not mounted yet.
Reinstall ZCS-88 with Next Generation Modules
When you are ready to reinstall ZCS-88, make sure the /opt/zimbra/ is now mounted.
If it's possible, rename your production /opt/zimbra/ directory.
mv /opt/zimbra/ mv /opt/zimbraSAVE/
Now, run the installer as the root user using the software option.
./install.sh -s *OR* ./install.sh --software
During the ZCS-88 Software Reinstall
Observe at the end of the software install, you will see the following
Running Post Installation Configuration: Software Installation complete! Operations logged to /tmp/install.log.SviTipi8
Now observe, you have two zimbra directories
drwxr-xr-x 30 root root 4096 Aug 24 14:29 zimbra drwxr-xr-x 30 root root 4096 Aug 17 13:20 zimbraSAVE
Since the software install had created the /opt/zimbra/ directory, lets move it out of the way
mv /opt/zimbra/ /opt/zimbraSOFTWARE_INSTALL/
Now, lets put our production directory back into place
mv /opt/zimbraSAVE/ /opt/zimbra/
After putting your production /opt/zimbra/ directory back into place, switch to the zimbra account, and start Zimbra
su - zimbra zmcontrol start Host test3.zextras.local Starting ldap...Done. Starting zmconfigd...Done. Starting logger...Done. Starting convertd...Done. Starting mailbox...Done. Starting memcached...Done. Starting proxy...Done. Starting amavis...Done. Starting antispam...Done. Starting antivirus...Done. Starting opendkim...Done. Starting snmp...Done. Starting spell...Done. Starting mta...Done. Starting stats...Done. Starting service webapp...Done. Starting zimbra webapp...Done. Starting zimbraAdmin webapp...Done. Starting zimlet webapp...Done. Starting imapd...Done.
Check the status
zmcontrol status Host test3.zextras.local amavis Running antispam Running antivirus Running convertd Running imapd Running ldap Running logger Running mailbox Running memcached Running mta Running opendkim Running proxy Running service webapp Running snmp Running spell Running stats Running zimbra webapp Running zimbraAdmin webapp Running zimlet webapp Running zmconfigd Running
Scenario I - Summary
Operating system needed to be installed, moving aside production Zimbra, and doing a software install, and placing production Zimbra directory back into place.
At this point, you should be up and running; however, as an option, you may want to check the user BLOBs, you can refer to this guide.
Scenario II - Mysql Server with corrupted mysql data that is beyond repair
You have been running ZCS-88 using the Next Generation Suite, and you have just discovered a lot of errors in the /opt/zimbra/mysql_error.log file, and the mysql service is not running
You even tried to export and reimport the mysql data to no avail.
Stop Zimbra Services
Move aside your old mysql directory
mv /opt/zimbra/db/data/ /opt/zimbra/db/OLD_data/ mkdir /opt/zimbra/db/data/
Start up mysql in a clean state
Start up the rest of the ZCS services
zmcontrol start zmcontrol status
The mailboxd service must be running in order to use the zxsuite command.
Run the external restore command to rebuild the mysql data
zxsuite backup doExternalRestore /opt/zimbra/backup/
Scenario II: Summary
If mysql service is not running because of corruption, simply move aside the data directory, and run the zmmyinit script to bring up mysql service into a new clean state. Followed up by running the doExternalRestore option.
Scenario III - Multi Server: Restore LDAP Service
Two Node Setup
Server 1 - LDAP + Proxy + MTA
Server 2 - Mailboxd (User mail store)
If LDAP services are needed to be restored, you can do the following.
Stop ldap and move the old database aside
ldap stop mv /opt/zimbra/data/ldap/mdb/db /opt/zimbra/data/ldap/mdb/OLD_db/ mkdir -p /opt/zimbra/data/ldap/mdb/db
Copy the backed up ldap data from Server 2
mkdir /tmp/ldap/ scp firstname.lastname@example.org:/opt/zimbra/backup/ldap/ldap_23_08_17#04_01_00.tar.gz /tmp/ldap/ chown -R zimbra:zimbra /tmp/ldap/
Extract the ldap.ldif file AND ldap-config file
tar -zxvf ldap_23_08_17#04_01_00.tar.gz ldap.ldif tar -zxvf ldap_23_08_17#04_01_00.tar.gz ldap-config.ldif
Restore the ldap-config
cd /tmp/ldap/ mv /opt/zimbra/data/ldap/config/ /opt/zimbra/data/ldap/OLDconfig/ mkdir /opt/zimbra/data/ldap/config/ /opt/zimbra/libexec/zmslapadd -c /tmp/ldap/ldap-config.ldif
Use zmslapadd to add the ldap data
/opt/zimbra/libexec/zmslapadd /tmp/ldap/ldap.ldif ldap start
List all labels
zmrestoreldap -lbs **Should see something like...** full-20170824.050005.077 full-20170823.050004.754 full-20170822.050004.779 full-20170821.050004.736 full-20170820.050005.331
Then use zmrestoreldap with the label name to restore the ldap data.
zmrestoreldap -lb full-20170824.050005.077
If ldap is not starting, use the long command to start ldap
sudo /opt/zimbra/libexec/zmslapd -l LOCAL0 -u zimbra -h 'ldap://ldap.example.com:389 ldapi:///' -F /opt/zimbra/data/ldap/config -d -4
The bottom portion of the output will usually provide a clue on why ldap is not starting.
Section III: Summary
Method A is to illustrate you can copy the backed up ldap from another node with Next Generation Module installed, extract the ldap data, and manually add ldap back into the ldap service.
Method B is to illustrate the old standard way of restoring ldap service using the zmrestoreldap command.
Next Generation modules are only installed on Zimbra nodes running mailboxd service; and by convention, nodes running mailboxd will need to remove zmbackup from cron.
However, in a multi node setup, Next Generation modules would not be installed on a node running ldap service. Therefore, the zmbackup command can still be in place in cron to save the ldap config.