sdani - 11:29 pm on Mar 26, 2009 (gmt 0)

>>>>>Ya, I have several backups online and here on disk. The probably is keeping them all current but the most recent one.

This is what I do (kinda plug to a software that I create, without naming it - but it works).

1. On my dedicated servers, I have scheduled tasks to export the databases to specific folders hourly & daily.

2. I then rsync and robocopy (depending on OS) all required folders, including db_dump, scripts, logs etc to a "tobackup/yyyymm" folder. Yes, this requires a little over 2 times of the local space on the dedicated server, but larger disks on servers are cheap (comapared to bandwidth).

3. Then I upload the "tobackup" folder to Amazon S3. Robocopy (or rsync) and upload to Amazon S3 is done in a single schedule job. This way, I have monthly backups available at S3, which are updated every morning. Once in 4-5 months, I login to my S3 account and delete older backups.

4. To backup the backup, I have another scheduled job, which runs on my home machine, and downloads the files from Amazon S3 to an external hard drive and then robocpy's that drive with another similar sized drive. This way, I have data on two local disks at home for past several months, and at Amazon S3 for past few months.

Very inexpensive, and it works. I have 1 day old data, available on a high speed network, which can be restored very fast.