Backup outline:
- Open the new backup drives and label them clearly Backup-A, Backup-B and Backup-C
- Take one power supply off site and keep two power supplies on site one in use and one somewhere accessible for rapid recovery from a power supply failure
- Swap drives weekly (or at least monthly)
- Two drives always remain on site during business hours
- One drive is always kept off site and never brought to the office
- On a weekly basis one drive is taken off site and the off-site drive is returned on the next business day
- The back-up is automated and starts at 10pm each night finishing between 4am and 8 am the following morning
- Backup speed is around 30 gigabytes per hour
- All users need to be logged out before 10pm
- Notify the helpdesk if a new disk (i.e. a replacement, not previously used) is added to the backup routine
Detailed procedure: assumes the use of 3 external hard drives labeled as above Backup-A, Backup-B and Backup-C
Week one Backup-A backs up all week (7 data blobs are created Saturday to Friday) then on Friday (before 10pm) the drive is swapped for Backup-B and Backup-A is taken off site on Monday Backup-C is returned to the office ready for the Friday changeover this way there is always 2 drives on site in case of drive failure and one drive always off site in case of a disaster.
Week two Backup-B backs up all week (7 data blobs are created Saturday to Friday) then on Friday (before 10pm) the drive is swapped for Backup-C and Backup-B is taken off site on Monday Backup-A is returned to the office ready for the Friday changeover this way there is always 2 drives on site in case of drive failure and one drive always off site in case of a disaster.
Week three Backup-C backs up all week (7 data blobs are created Saturday to Friday) then on Friday (before 10pm) the drive is swapped for Backup-A and Backup-C is taken off site on Monday Backup-B is returned to the office ready for the Friday changeover this way there is always 2 drives on site in case of drive failure and one drive always off site in case of a disaster.
Then repeat
