John P . Looney stated the following on Wed, Feb 21, 2001 at 03:21:34PM +0000 :
> I didn't. What I've running at the moment is ipvs & mon & heartbeat
> setup to do clustering for mysql, http, postfix and courier imap. Works
> well - but there is a ~2 second gap between mon noticing that a service is
> down, and changing the routing tables for it. It nats everything out of
> the cluster, so it looks like it's all one big machine.
>> Heartbeat's just used for a fail over router. Problem there being that
> you lose the nat tables - so if your router dies, all your connections
> drop, and it's a second or two before you can reconnect to anything. Not
> brillant.
Need to go to something like a foundry serveriron for that kind of
seemless failover.
> > You need fibre channel for the 16 nodes - mucho deneros, else you can use
> > shared scsi and only two nodes. But you can cluster applications, we setup
> > MySql and Apache, and brutally shut off machines for the laugh, and it was
> > almost seamless as far as the clients were concerned - 20 seconds as far as
> > I can remember - but this is adjustable - and better than MSCS can do.
>> How does it all work ? What does it use for shared storage (which is my
> big problem at the moment. NetApp clusters are expensive, as are dual
> attach SCSI enclosures.
Have you tried using GFS,
--
"Razor blades. The currency of the New Economy."
-- Joel Spolsky http://joel.editthispage.com
Maintained by the ILUG website team. The aim of Linux.ie is to
support and help commercial and private users of Linux in Ireland. You can
display ILUG news in your own webpages, read backend
information to find out how. Networking services kindly provided by HEAnet, server kindly donated by
Dell. Linux is a trademark of Linus Torvalds,
used with permission. No penguins were harmed in the production or maintenance
of this highly praised website. Looking for the
Indian Linux Users' Group? Try here. If you've read all this and aren't a lawyer: you should be!
