[GNHLUG] MerriLUG Nashua, Thur 19 Jun, MySQL: The Whys, Whats, and Watch-outs

[Mark Komarinski]

First, sorry for top-posting.  I'm on my new blackberry and still getting used to it (anyone know how to do a proper bottom-post?)

We were going to go active/active with our MySQL system, but the autoincrement thing really got us nervous.  That and I was able to get both masters out of sync pretty easily.  Once that happened it was almost impossibble to figure out which was 'correct'. Our MySQL install has about 80 databases and 120 GB of data, so getting that out of sync would be a mess to clean up.

DRBD works great and between that and heartbeat, we have a pretty robust setup.  Not active/active but MySQL can failover and be available in about 30 seconds.

-Mark

-----Original Message-----
From: "Flaherty, Patrick" <pflaherty at wsi.com>

Date: Mon, 23 Jun 2008 17:31:50 
To:<gnhlug-discuss at mail.gnhlug.org>
Subject: RE: [GNHLUG] MerriLUG Nashua, Thur 19 Jun, MySQL: The Whys, Whats,
	and Watch-outs


> I've made my slides available here:
> 
> 	http://docs.google.com/Presentation?id=dcwc3b2p_63ck9zndhh

I didn't go to the meeting, but I saw the slides mentioned DRBD, which
I've been using lately to redundantify everything (nagios, mysql,
cacti). I could do a presentation on it if anyone is interested.

As a side note, is anyone doing mysql active/active clusters? The whole
primarykey even/odd thing seemed really dirty to me.

Patrick

_______________________________________________
gnhlug-discuss mailing list
gnhlug-discuss at mail.gnhlug.org
http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/