The Bits 2.0

Data loss is one of the leading cause of hair loss to IT administrators and anemia to the company’s treasury.

Here’s a quick and simple but effective guide to preventing data loss disasters for a typical webserver (and probably preventing you from going bald):

1. Backup Often and Many

Previously, I set my server to backup databases nightly and web files weekly. The whole disk backup will be carried out by my service provider so I never bothered about it. Backup tar balls are sent to a backup file server in Malaysia.

After the hard disk crash disaster on March 24, I’ve radically changed my disaster prevention plan. Databases are tar balled twice and sent to the Malaysian backup server and a United States backup server daily, both during during low traffic hours. Web files are tar balled and rsync-ed with the Malaysian backup server twice a week.

2. Separate Backup Locations

I’ve seen people storing their “backups” on different partitions. That is totally wrong! Here’s a list of how to store backups, from worst to best:

a. Store on the same partition
b. Store on another partition
c. Store on a physically different disk
d. Store on an onsite backup server/NAS
e. Store on an offsite backup server which is in the same city as your server
d. Store on an offsite backup facility located on the west coast when your server is located on the east coast of USA.
e. Store on multiple geographically dispersed backup facilities…but hopefully a reliable datacenter is housing your backup server.

Why store on geographically dispersed locations? Because an earthquake or hurricane could take down an entire state, and your data, with it.

3. Ask the Professionals

If your datas are too important and valuable to lose, contact professional data backup companies and they will give you the best possible ways to prevent data loss disasters. But be prepared for hefty price tags.

Trackback URI | Comments RSS