[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: N-Way Multi-Master when one server down

Ok, attached (to protect from being as easily searched from the web crawlers) is my slapd.conf that has been scrubbed of some sensitive information. I'm using the slapd.conf version versus the cn=config as I had easier success with it.

The second server's syncrepl configurations are the exact same. I can include that server too if anyone thinks that is where the problem lay lie.


Attachment: slapd.conf.clean
Description: Binary data

On Jan 15, 2008, at 11:25 PM, Howard Chu wrote:

Chris G. Sellers wrote:
I'm noticing that with N-Way Multi-master servers, if one server is
down or crashed, the updates don't always get to this server once it's
back up. (e..g updates are not queued and when the connection re-
establishes it does not flush out that queue)

Then something is broken.

Do I understand this to be correct?  Can you adjust the retry time
settings to help mitigate this?

Retry settings should make no difference. When a downed server comes back up, it does a refresh operation with the other server to bring itself up to date. You should post your configs here again.
-- Howard Chu
Chief Architect, Symas Corp. http://www.symas.com
Director, Highland Sun http://highlandsun.com/hyc/
Chief Architect, OpenLDAP http://www.openldap.org/project/

Chris G. Sellers			|	NITLE Technology
734.661.2318			|	chris.sellers@nitle.org
AIM: imthewherd			|	GTalk: cgseller@gmail.com