[SOLVED]CRITICAL! LDAP primary MDB database is 90% full in Zimbra

Zimbra logo

You may get this type errors as email in the zimbra admin email account. This was related with the zimbra 8.x changes in ldap db configuration. You may probably see this issue during a zimbra migration or restore task
First check the following values in your zimbra

#su - zimbra
#zmlocalconfig -s ldap_db_maxsize
ldap_db_maxsize = 85899345920
#zmlocalconfig -s ldap_accesslog_maxsize
ldap_accesslog_maxsize = 85899345920

You can see the db size and accesslog size is same that is 80 GB , Now please setup the size as follows,

# zmlocalconfig -e ldap_db_maxsize=85899345920

# zmlocalconfig -e ldap_accesslog_maxsize=10737418240

Give some time for applying the changes

# sleep 90

Now stop slapd

# ldap stop

Go to database directory

# cd /opt/zimbra/data/ldap/mdb/db

Take a backup of ldap database

# /opt/zimbra/libexec/zmslapcat /opt/zimbra/data/ldap/mdb/db

Rename the original db to .old

# mv data.mdb data.mdb.old

Recreate database with new size and restore from backup

# /opt/zimbra/openldap/sbin/slapadd -q -b "" -F /opt/zimbra/data/ldap/config -l /opt/zimbra/data/ldap/mdb/db/ldap.bak

Now start slapd

# ldap start

Tar the original if you need

# gzip data.mdb.old

The above will fix your issue

How to Fix hardware clock issue from a Xen VM

You can run an independent clock for your XEN VM . That will fix the hardware clock problem, unless you don’t have access to host node. It is always best to keep the hardware clock of Dom0 with synchronization . But if your server provider won’t give support for you , please set it as follows,
Setup the wall clock in /proc

echo 1 > /proc/sys/xen/independent_wallclock

And to keep the setting between reboots, just add it to the /etc/sysctl.conf

xen.independent_wallclock = 1

This will fix the issue.