<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
The web machine uses a PHP shopping cart running atop MySQL, all of the
main and important information is kept as a part of the MySQL DB. So
the main thing I would want to mirror is the application and the DB.
I'm a surplus dealer, so hardware isn't a big issue, I have extra
servers that can be up in a couple of hours without any big layout of
money.<br>
I'm just looking for the best way to keep down time to a minimum and as
always to KISS, Keep It Simple Stupid. Simple to manage, simple to
maintain...<br>
<br>
John<br>
<br>
<br>
<br>
Les Mikesell wrote:
<blockquote cite="mid:4AB3CB50.1090007@gmail.com" type="cite">
<pre wrap="">John Plemons wrote:
</pre>
<blockquote type="cite">
<pre wrap="">I want to mirror a Centos box, not having done this before, I'm looking
for guidance.
What is the best way to perform this task, I have a running webserver, a
single external IP address to the website, and wanted to protect myself
should the server go down. My though would be a mirror of that
webserver, running in background, if the main server dies, then just
move the mirror in place, change the IP and go on... Limiting down time...
</pre>
</blockquote>
<pre wrap=""><!---->
There are several approaches, depending on how you want to trade expense
and effort against possible downtime - and whether your site is
approaching the point where you need a load-balanced farm of servers and
possibly redundant sites.
The easiest plan with the least to go wrong is probably to use RAID1
mirrored/swappable drives with a spare chassis, and also do nightly
backups. The most likely failure will be a disk drive, handled
transparently by the mirror until you can replace it. Next most likely
would be the power supply/motherboard which you handle by swapping the
disks to your spare with a small amount of downtime. Next would be an
operator or software error that erases or corrupts your disks. For that
you have to restore from your last backup with a much longer downtime.
You can keep the spare server in sync with DRBD and fail over
automatically with heartbeat at the expense of more complexity but you
still need backups for the error scenaro.
</pre>
</blockquote>
</body>
</html>