Create
cancel
Showing results for 
Search instead for 
Did you mean: 
Sign up Log in

Crowd Multiple LDAP servers

A February 18, 2018

I have 2 LDAP servers Primary (currently configured via ldap: chemar URL field in connector, but want to add 2nd server (LDAP Slave). Should I specify 

URL: ldap://ds1:389/ ldap://ds2:389/

instead of:

URL: ldap://ds1:389/

3 answers

1 accepted

1 vote
Answer accepted
Marcin Kempa
Atlassian Team
Atlassian Team members are employees working across the company in a wide variety of roles.
February 19, 2018

You may also give a try directory failover that Crowd offers, as described here.

Or you can go for the load balancer method as suggested by @Craig Castle-Mead.

 

Hope that helps,

Marcin Kempa

1 vote
A February 19, 2018

Crows is based on java naming service API, the URL field may contain multiple servers:

https://docs.oracle.com/javase/jndi/tutorial/ldap/misc/url.html

 

 

But it's not documented by crowd team.

Craig Castle-Mead
Rising Star
Rising Star
Rising Stars are recognized for providing high-quality answers to other users. Rising Stars receive a certificate of achievement and are on the path to becoming Community Leaders.
March 22, 2018

Thanks @A - was just able to get time to fully test this out and it's simplified things significantly (no need for external haproxy/lb/etc.

Used the URL format

"ldap://10.0.0.1:3268/ ldap://10.0.0.2:3268/"

Whatever the connection timeout value is, it'll check against each instance, and if there's no response, moves down the list

CCM

0 votes
Craig Castle-Mead
Rising Star
Rising Star
Rising Stars are recognized for providing high-quality answers to other users. Rising Stars receive a certificate of achievement and are on the path to becoming Community Leaders.
February 19, 2018

I’m quite certain that the UI won’t even let you save with the double LDAP server string in the field.

I’d suggest looking at load balancing your ldap servers outside of Crowd itself (f5, AWS elb, haproxy), have Crowd point to that device and then it has the smarts to know which server to route the traffic to.

If the two servers have replication delay, you may notice some weird behavior if/when you need to fail over. Hopefully this doesn’t happen too frequently, but I’d definitely suggest testing it out on a non-prod install, or even just a new Crowd directory that isn’t linked to any of your apps. Force a failover, see what happens the next sync, fail back, test again. I’d defnintely suggest against using a round robin approach (lowers the risk of constantly adding/removing groups of the two ldaps are out of sync), only use the secondary when you have to.

 

CCM

Suggest an answer

Log in or Sign up to answer
TAGS
AUG Leaders

Atlassian Community Events