Howto make your server highly available

Setting up a highly available NFS server (5 pages)
http://www.howtoforge.com/high_availability_nfs_drbd_heartbeat

Mirror Your Web Site With rsync (2 pages)
http://www.howtoforge.com/mirroring_with_rsync

How To Set Up A Loadbalanced High-Availability Apache Cluster (4 pages)
http://www.howtoforge.com/high_availability_loadbalanced_apache_cluster

How To Set Up A Load-Balanced MySQL Cluster (8 pages)
http://www.howtoforge.com/loadbalanced_mysql_cluster_debian

How To Set Up Database Replication In MySQL (2 pages)
http://www.howtoforge.com/mysql_database_replication

Uptime

I was always more interested in 'computing', rather than 'computers'. Its just a pity it took me so long to get my head around programming.

The topics that I found most intriguing, and thus in need of further exploration were:

Reliability
Scalability
Redundancy
Fail-over
Fault-tolerance
Load-Balancing
High Availability (HA)
Clustering
Beowulf
High Performance Computing (HPC)

Most - if not all - of these, are likely to be understood and embraced by the Linux / UNIX and Mac OS/X community, whilst still being completely alien to Windows Administrators and users =D

Now there are those who will state that 90% of the computers in the world run some form of Microsoft OS. That's all well and good, but how many Ford Escorts do you see in a Formula One Grand Prix? Popularity is not everything, people! Microsoft started out as the industry's cheap and cheerful alternative to a real os, and succeeded through aggressive marketing and somewhat questionable business ethics (if any?). Indeed, I'm not trying to convert any Windows users off their platform, they can keep it if it makes them happy.

Instead, I'm eventually going to discuss how I apply the principles learned from the above concepts in running my own cluster of machines. All I want to show you right now is the current state of the 3 machines currently running in the cluster.

As for the servers in the cluster though, they are as follows:
The Mac Mini (physical), which is hosting one of the BlueQuartz servers (virtual).
The other two systems are brand new HP ProLiant ML115's (2 x physical).


osx: $ uptime
21:51 up 24 days, 11:49, 2 users, load averages: 0.21 0.33 0.37

bq $ uptime
23:32:25 up 24 days, 12:25, 1 user, load average: 0.04, 0.06, 0.02
(hey, my virtual machine has been running longer than the host system, osx ! trippy!!)

ubuntu $ uptime
4:51:25 up 23 days, 23:54, 1 user, load average: 0.00, 0.01, 0.00

bq2:~$ uptime
22:05:27 up 7 days, 12:44, 3 users, load average: 0.96, 0.20, 0.06


As you can see, they are also largely unused. This should change, as I migrate applications from the notebook (where I do most of my development), onto the cluster so I can see my software projects running as a 'distributed application'.

p.s. I am also aware that I am going to maybe shut one of the inactive servers down for a memory upgrade (the other two have already had theirs), so the figures will soon be getting reset to zero.