Hello Beautiful People !!!

[Madison Kelly]

Welcome to the list!

A topic I know others will raise; Please send email in plain-text 
format, not HTML. Several people here use text-only mail readers.

Dev Guy wrote:
> Wow my first day on the thread as a new member to the Linux group and 
> we're debating the creator and religion!

GTALUG is a community. :)

> Hello everyone, great to be in the company of open minded passionate 
> people! I've joined this group to develop a deeper understanding of 
> Linux. I am a Slackware oldie but newbee, and currently CentOS has 
> caught my attention!

Welcome back!

> I am here out of selfish reason to learn how to be a better Linux 
> administrator with CentOS. I plan to use these skills to manage my 
> website down the road.

Ultimately we're all here for the "selfish" reason of wanting help.

> What I need to figure out is how to use the capabilities of 
> virtulization or maybe it's clustering to use multiplte servers at the 
> software level to create a load-balanced environment? For example, 
> creating a online group that will require more server overtime to 
> accomodate the demand of increased traffic and increased storage. How 
> could one go about getting this done?

You need to understand the differences between virtual servers and load 
balancing.

Virtual servers:

These are machine installed inside a virtualized environment like Xen, 
VMWare and so on. The major benefit of this is that the entire "machine" 
is really just a file. You can very easily back this up, move this 
between hardware machines and so forth. It is primarily a convenience 
thing. For example, when your underlying hardware no longer provides 
fast enough performance, you can build a new machine with new hardware 
and an updated OS and, when you are ready, simply copy the virtual 
machine's file over and start it up. No need to worry about all the 
wierd and esoteric changes you've made over the years.

Clustering:

This is done for one of two reasons; the first being for spreading the 
load out among multiple machines and the second is ensure your services 
are provided even when a given server blows up with little to no 
downtime. In either case, multiple virtual machines on one server will 
provide no such benefit.

> I plan to go with a host and start off with a single server and later 
> add server as required. Is there a way for me to scale easily without 
> the use of handware other than the servers?

Are you planning to run a custom written application? If so, it needs to 
be designed for some form of clustering and/or fail-over. If you are 
using a canned application, you will need to ensure that it supports the 
same or that it can be supported by something like heartbeat/HA-Linux.

> Is there any opensource software that will allow me to do this using a 
> GUI tool on CentOS in particual?

Again, depends on what you want your cluster to do. In some cases 
though, yes. Look at: http:///ha-linux.org

> Thanks for your assistance in this matter and I hope I can contribute to 
> the group as my skills develop.
> 
> Kind Regards,
> Rajinder Yadav

Something to keep in mind; Clustering introduces many new points of 
failure. They are something that need to be carefully studying and 
planned for. Specifically, is the added complexity and risk sufficiently 
offset by some benefit? Ask yourself; "Would a second spare server that 
periodically copies the changed files from the main server act as a 
sufficient spare?". Ask other similar questions.

In all cases, choose the simplest solution that addresses your needs and 
you will be a happy admin. :)

Madi
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists