Listener/Notifier replication


#1

Hello!

Looks like the replication Master/Slave is broken since the latest update.
Both servers were updated up to 4.2-2 release version.

In /var/log/univention/notifier.log file on Master I see the following events:

07.09.17 11:17:21.693  TRANSFILE   ( ERROR   ) : 7 failed, close connection to listener
07.09.17 11:17:55.102  TRANSFILE   ( ERROR   ) : 7 failed, close connection to listener
07.09.17 11:18:32.939  TRANSFILE   ( ERROR   ) : 7 failed, close connection to listener
07.09.17 11:19:08.450  TRANSFILE   ( ERROR   ) : 7 failed, close connection to listener

Meanwhile, in /var/log/univention/listener.log file on Slave I have these events:

07.09.17 18:18:38.180  DEBUG_INIT
UNIVENTION_DEBUG_BEGIN  : uldap.__open host=ldap.example.com port=7389 base=dc=example,dc=com
UNIVENTION_DEBUG_END    : uldap.__open host=ldap.example.com port=7389 base=dc=example,dc=com
07.09.17 18:19:43.539  LISTENER    ( ERROR   ) : connection to notifier was closed
07.09.17 18:19:43.539  LISTENER    ( ERROR   ) : failed to recv result
07.09.17 18:19:43.539  LISTENER    ( ERROR   ) : listener: 1

File /var/lib/univention-directory-listener/notifier_id on Slave contains number 2723249
The same file ( /var/lib/univention-directory-listener/notifier_id ) on Master has number 2723419

Connection between servers is not prohibited and simple tests:

From Slave
nc -vz ldap.example.com 7389
Connection to ldap.example.com 7389 port [tcp/] succeeded!
nc -vz ldap.example.com 6669
Connection to ldap.example.com 6669 port [tcp/
] succeeded!

From Master
nc -vz ldap-slave0.example.com 7389
Connection to ldap-slave0.example.com 7389 port [tcp/*] succeeded!

show that replication can be performed.

Unfortunately, we observe the errors above on Master and Slave servers.

Could you give me the direction which I should dig in for resolving the issue?

Thanks in advance.


#2

Good Morning sly_roar,

  • The result of /var/lib/univention-directory-listener/notifier_id for both systems should be the same. In your case they are not.
  • Check if the replication was complete by running the following command on the master and slave.
# /usr/lib/nagios/plugins/check_univention_replication 

Should there be an error or warning message then,

  • set the debug level on the slave to 4 and restart the service.
# ucr set listener/debug/level='4'
# /etc/init.d/univention-directory-listener restart
  • You can now take a look at the listener.log file
# less /var/log/univention/listener.log 

For further help, take a look at this article:

My regards,
Anna Takang