Month: November 2004

Coming to a web server near you: the ultimate in low(ish) cost high availability web serving. The concept is to use another dedicated server from the same company, but physically located in a different country. Load up all the relevant files, install SQL Server and set up replication between the two sites at hourly intervals. If the primary goes down, simply give out the url of the backup server.

Seems a good idea, will have to wait to see how it works in practise.


Having had so many things go down unexpectedly recently I’m still in backup paranoia mode: saving files regularly whilst working on them; using my external hard drive to copy all my working folders and files; using SQL server’s DTS scheduling to download stored procedures and databases that aren’t copied using replication. Of course, as long as I’m doing all of this the hardware will work fine….


I haven’t been able to bring myself to mention the fourth disaster. To a certain extent it was a planned outage. We did the backup, notified users of down time and then proceeded to rebuild the server. The server came up, SQL was installed, the SQL backup was restored … not!

Did anyone ever mention that backups are no good unless verified? How often in the heat of a ‘situation’ does anyone ever stop to verify that backup. Don’t you just hate computers sometimes?

Of course hindsight is wonderful. Next time there will be multiple backups, verification and all sorts of other paranoid checks. But that’s next time…


I have two broadband accounts at different locations, both required to have a fixed IP address. One, provided by Eclipse Internet allowed this by logging on to my account and ticking a check box.

BT have so far:

1. Sent me a new broadband modem (not required).

2. Scheduled a time for the upgrade (a couple of weeks later).

3. Taken the connection down for the entire day to change a non connection specific item, ie. the IP address.

Nuff said!