GitLab melts down after wrong directory deleted, backups fail

theregister.co.uk

12 points·dsaw··7 comments

(obviously?) a dupe https://news.ycombinator.com/item?id=13537052

No new information in the article itself. But articles in The Register usually have some good comments. E.g. this article there has 141 of them, at least some of which are good. :)

I've never actually experienced a work situation where needed backups did work. I fear this is the norm, not an aberration.

Really? What have you used for backups?

You may be surprised at the average success rate for Backup Exec 2010-2012 across a common enterprise environment. It is entirely normal to look at failures and state "it's working 50% of the time, we should be happy". Their own support team would look at that sort of log and say things like "you seem to be having really good luck with our product".

I wish I could remember where I first heard this quip: "if you don't test your backups, you don't have backups, you have dreams".

Articles like these crop up periodically on HN. They're very useful because they reiterate the point about the necessity of testing the backups.

Jesus, that's terrifying.

Since moving almost all servers I use to VM's, my backup restore success rate is very high. We generally take a VHD image at least once a week, and then daily processes (maybe even hourly for databases) will run inside the OS itself.

Reply on news.ycombinator.com