[Date Prev][Date Next]
Re: Fault-tolerance for master OpenLDAP server
- To: firstname.lastname@example.org
- Subject: Re: Fault-tolerance for master OpenLDAP server
- From: "Samuel Tran" <email@example.com>
- Date: Wed, 31 Aug 2005 21:55:29 -0400 (EDT)
- Cc: openldap-software@OpenLDAP.org
- Importance: Normal
- In-reply-to: <1770700965-1125538060-cardhu_blackberry.rim.net-19181-@engine07>
- References: <1770700965-1125538060-cardhu_blackberry.rim.net-19181-@engine07>
- User-agent: SquirrelMail/1.4.4
> You're thinking about this wrong. You want one master server and two
> loadbalanced slave servers, or loadbalance the master/slave for requests
> and only do updates on the master. It was designed to be pretty fault
> taulerant, and drdb kinda sucks (takes forever).
> There are plenty of loadbalancers out there, but I've had good experences
> with zeus's zxtm, foundry server irons, and heard good things about
I have already a setup that includes one master and several slaves that
are load balanced using Keepalived (http://www.keepalived.org). This has
served me very well for a year.
What I want now is to have a second master that would take over if the
primary master goes down, i.e. reduce the time it would take to have a new
master fully operational.
Thanks for your input on DRBD anyway.