←back to thread

4 points atomicnature | 2 comments | | HN request time: 0s | source

Recently, we lost some data during migrating our servers due to a missing backup.

We thought we had something backed up - but was not really the case.

We have multiple databases and apps - each having its own data store often.

How do you usually deal with server backups? What has worked for you and what has not?

1. codegeek ◴[] No.44494066[source]
Some rules for backups that you must follow:

1. Backups must be taken offsite on a separate server (obvious but surprisingly some people miss this)

2. Backups must be tested frequently. If you cannot test a backup, you don't have a backup.

3. Frequency depends on your criticality of data, your contract/SLA with your customer etc. Ideally, you should be able to have Point-in-time-Restore (PTR) going back to certain number of hours/days/weeks

4. Make sure to have notifications for backup failures. If a backup failed, you must be notified to correct it manually.

5. Bonus: Have a backup reconciliation script that runs additionally to recon all backups for a certain period.

replies(1): >>44497567 #
2. chistev ◴[] No.44497567[source]
> Backups must be taken offsite on a separate server (obvious but surprisingly some people miss this)

Yea, you can't call it a back up if it's still in the same place. Lol