Re: [Nolug] Quick question on clustering...

Joey Kelly <joey@joeykelly.net> writes:

> Many times, the web server is physically separated from the database, so that
> one machine doesn't have to do the work of processing requests from the user
> (what link did the user click on, etc.), rendering pages, and hosting the
> database application. If the database work is shuffled off to a separate

Right. And this is generally a good idea to do in a production
environment. Of course this is really network load-balancing rather
than "clustering" per-se. But based on the context, I figured that was
what Craig meant by the question.

> What about a cluster of database servers? Perhaps Ron can jump in here, but
> if most of your database requests are selects or reads, then you can probably
> replicate your data at intervals to several slave machines, while your writes
> to the database would be sent to the master database, and these changes would
> propagate to the slaves during scheduled replications.

all the major databases can do this -- mysql and postgres included.
I've implemented replication on mysql. If your application is written
correctly, you should be able to fail over gracefully. DNS is often
used in this scenario to provide a "virtual hostname". There's more
than one way to do it though...

> Ron was telling me the other day about SAN (storage-attached networks), where
> the database servers were networked together not by ethernet, but by scsi or
> some other data connection to the storage media. This layout is kind of like
> a bicycle tire, where the storage media is at the hub, and the database
> servers are on the perimeter. They all share the same hard drive cluster
> (RAID) for their reads and writes, therefore eliminating the need for the
> staggered scheme detailed above. Replication in this instance would only be
> needed for backup purposes.

Yeah. I'm doing a SAN here. SAN's are generally not SCSI but fiber
channel. A SAN involves attaching your FC adapters to a SAN
switch. (I've not seen a SCSI SAN switch, but then I'd never looked to
see if such a thing did/could exist) The switch is cabled to the
drives as well. It's roughly analogous to a LAN, hence the name. I'm
not a DB admin, but I don't think you'd have multiple database
instances writing to the same database though. The locking could get
pretty complex.

With clustering software you can do this with direct attached storage.
One box has to have hold of the drives though -- you can't read write
simultaneously. The software in conjunction with a private network
and some sort of detection mechanism determines if a cluster node has
failed and switches the mountpoint (and services if this is all
implemented properly) to the other box.

-- 
Scott Harney<scotth@scottharney.com>
"...and one script to rule them all."
gpg key fingerprint=7125 0BD3 8EC4 08D7 321D CEE9 F024 7DA6 0BC7 94E5

___________________
Nolug mailing list
nolug@nolug.org