[Date Prev][Date Next]
Replicating a translucent database
- To: email@example.com
- Subject: Replicating a translucent database
- From: David Arroyo <firstname.lastname@example.org>
- Date: Thu, 22 Mar 2012 23:20:44 -0400
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:date:message-id:subject:from:to:content-type; bh=o2gF2YxKcCeEz57fZ1VM6pbRtH9ZqGAOMHmK4n6FedQ=; b=stBhGHQh0zw+6y8L33Npre+xm6j9nS79IqGE9s8Mwmw7Uh1xyRoFxj0DAWq3mIAFoK JLP34WFWAWYIH8LnDNefQHfho/bgGyNRjjxOSxw4ZpAkaRvCP2AZr/jcHVgtQDULE/ne W8Ns7Zg0TaQBezSD8EGxUdV8oOnk0N4kdX8XPcqKrBwlbgkDvkRjV4/BB1K6YNKIcun4 aDdk4QGAjlTt3J8LsR+Ah9m0lg32957A9LPGuPTvP1dbW4IhkdYtdcaIRvf5uiYSfF47 Nlg3cfsZ9NFn3UXPoWXjJl4gI86GxeUXm5cOZrY8RTX8Plt1Ton9Z/p1EZP1WrEJcliN 7zOA==
Does anyone have experience with replicating the "local" part of a
translucent proxy? I've gotten into the habit of running redundant
pairs for everything. Since translucent uses back-ldap, which sets
lastmod off, traditional replication via syncrepl doesn't work. Right
now I'm only supporting a directory of about 2000 engineers, so it is
feasible to use some combination of slapcat and ldapadd/ldapmodify
magic to update a secondary server.
I could also slapcat the translucent database and take down the
secondary server while the database is rebuilt. One idea I had is to
copy the local translucent database to a separate (non-proxy)server
every N hours, and have any slaves replicate off of that. However, due
to schema checking I cannot simply copy a translucent db to a
non-translucent one due to all the glue and missing attributes.
I've found the translucent overlay (and openLDAP's flexibility in
general) to be a lifesaver. If anyone has any ideas or experience with
this particular challenge, I'd love to hear about it.