[CentOS] Webserver redundancy - help

Wayne wayne at nightsol.net
Thu Mar 23 13:57:54 UTC 2006


Hey Guys,

I have 3 webservers rented with a datacentre. Each of them provides the same
webservices to different clients.
So clients A,B,C would log onto webserver 1.
Clients D,E,F log onto webserver 2.
Clients XYZ, log onto webserver 3.
Each of the webservers provide the same services using apache, php & mysql.
Data is collected from an application logged to sql and the users can view /
manipulate that data over the web.

My problem is there is no redundancy. If webserver 1 goes down then clients
A,B,C have no services.
I need to add in redundancy. I can't at the moment afford a 4th server.
So I need to make do with whats currently there.

Were talking about 100megs of new data on each server per day.
Would something like mysql replication be the best bet? Anybody have any
better ideas? 
Correct me if Im wrong but under replication wouldn't one server be a master
and all others be slaves?
So if I made webserver 1 the master then all clients would have to log into
that one if they want to manipulate the data and if it goes down them Im
screwed on all clients?

I know I need a better setup just wondering whats the best way to go about
it...

Thanks,
 Wayne





More information about the CentOS mailing list