He is on point: the hardware RAID controllers will only react if a drive is signaling something wrong. But if the drive is fine but data changed somehow the controller won't respond with anything. That's why ZFS is better for data integrity. Bit rot does exist however at the consumer level is not relevant because we run into the numbers of probabilities, the MTBF numbers the drive manufacturers specify: maybe for a single HDD this won't happen but when you have 100.000 HDDs those probabilities are just waiting to happen. That's why an SSD is strongly recommended when you build a NAS/ZFS setup for storing the journals. But all these are known facts he just wants to point them out. I am sure someone is already working right now to implement those policies of data protection/redundancy/integrity right now, because they are a necessity and a major selling point for data storage providers.
hardware raid is dead says Wendell from level 1 tech.
- Thread starter gingerbee
- Start date
Marzipan
Well-known member
anyone remember how like, 15 years ago we were so excited for death_hawk and his 100GB project? LoL!
now when these gpu raid cards come down from enterprise pricing we can set some blazing records.
