Using Crowd with redundant LDAP servers

[e]

Hi everybody,

in our company we have three LDAP-servers which point to one AD. This redundancy enables to use update/restart one of them, when the others are still online.

My question is now, do I encounter any problems when i configure three LDAP connectors in Atlassian Crowd which point to he same AD?

kind regards,

Andreas

[Tiago Comasseto · Accepted solution]

Although it's not officially supported, I've seem cases where customers have two identical LDAP servers and a load balancer in front of them. In this scenario Crowd could be configured pointing to the IP of the load balancer, instead of the actual servers.

Cheers

[Tomasz Kontusz]

You can add multiple LDAP addresses for one directory:

In Crowd 2: "ldaps://1.1.1.1:636 ldaps://2.2.2.2:636"

In Crowd 3: "1.1.1.1:636 ldaps://2.2.2.2" (Crowd will just concatenate "ldaps://" + address + ":636").

[Tomasz Kontusz]

In the end this solution broke too. Setting up a simple load balancer, as in the accepted answer, works well.

[Mike Monette]

...But why? Why do you guys put yourself into this position? It can be done......so......easily. Why not support it?

Using things like ucarp/relayd/carp/pacemaker and point crowd at the virtual shared IP works just fine...So why not allow us to enter 2 ldap servers in crowd to be used for redundancy? I don't get it.

[BernardoA]

Basically you can create several different LDAP directories but not for redundancy purposes.

[Renjith Pillai]

Connecting to replicated/redundant LDAP servers is not something which is is supported - https://confluence.atlassian.com/display/JIRA/User+Management+Limitations+and+Recommendations (that article is for JIRA and I guess it is the same for Crowd as JIRA has embedded crowd inside it)

Also - https://jira.atlassian.com/browse/CWD-422