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

Re: Documentation volunteer (was: Cluster replication: multiple masters)



I've tried the multimaster code in the 2.1.x series.  It's not "official,"
and you need to enable it by #define-ing the MULTIMASTER def in
includes/portable.h before running make to enable it.

My take so far is it works well.  You need each server set up like a
master (running a slapd and slurpd), and each configured to replicate to
all the others.

It's a good strategy even behind a load balancer... but I think you'll
need to set up your load balancer to allow direct IP access to each
server, as well as having the virtual IP address to do the dynamic
redirects.
-Alan

Derek Simkowiak said:
>
> 	So (with some help, no less) I managed to find the web-browsable
> list archives, cleverly hidden under the heading of "List Archives" on
> the lists web page.
>
> 	Searching for "multiple masters" gives a bunch of very old
> messages (95% more than a year old, many over two years old) mostly
> about some "experimental" code in the 2.0.8 CVS tree.
>
> 	My first question is, "huh?".  Why would multi-master replication
> require any code whatsoever?  It seems like you'd just configure slurpd
> as per my last message in this thread.  If someone familiar with this
> 'experimental' code could comment, I would be very grateful.
>
> 	The second question is, what is the current status?  Is anybody
> using this?  And again, why would this be anything more than having
> slurpd feed the log to multiple replicas in both directions?
>
> 	I would be happy to write an up-to-date mini-HOWTO on my
> experience, so if anyone knows anything about this (or has examples of
> where it's deployed), please reply.
>
>
> Thank You,
> Derek Simkowiak


===========
Alan Sparks, UNIX/Linux Systems Administrator    <asparks@doublesparks.net>