This is the most infamous one Fatal: number of used blobs is larger than number of available blobs! · Issue #3268 · restic/restic · GitHub
(And subsequent comment about “we only fixed files that were reported by support tickets”… Production storage service my ass. “if you continue seeing a problem let us know” – how about implement monitoring and verification?)
This shit (returning bad data to customers) should have been impossible on that product maturity stage.
(This post summarizes pretty well other shit backblaze does without batting an eye: Frankly, I’ve never seen a good take from Backblaze once I got past their market... | Hacker News (and their rebuttal on the lack of zero-knowledge encryption is full of shit – you cannot restore encrypted data from backblaze business backup service without giving them your keys)).
Just don’t use Backblaze. Don’t cheap out on your data storage quality and integrity. You get what you pay.
(I can also pour a bucket of shit on ProtoCase that manufactures their storage pods, and its subsidiaries – 45Drives and 45HomeLab… but that’s for another day)
To stay somewhat on topic – the horrific swarm of nonsense companies like Backblaze, iDrive, pCloud, and other “budget” participants in the race to the bottom makes Storj’s quality, reliability, attention to details, and how fast they do fix reported failures stand out even more; vast majority of the complaints about storj are all cosmetic and pretty much irrelevant in the grand scheme of things, and I’m yet to see data integrity issue. Keep up the good job folks. (See, my buckets of shit sometime carry honey. Rarely, but it happens)