Q about Minute 11:10 of "FreeNAS 9 Software Tutorial & Overview One" Video

Level1Techs INBOX.deb
1

Video Link: https://teksyndicate.com/videos/freenas-9-software-tutorial-overview-one

Q about Minute 11:10: Can that happen 'Data there but corrupted' even with a Drobo NAS HW/SW proprietary product?

Google says yes, if the volume's directory structure becomes corrupt, ur arsed.

PS: I already asked the same Q in YT comments, and the comment on the video page on teksyndicate. But in the end of the video u mentioned the video so I thought I post it here as well.

Thanks in advance for answering this question whether or not this (FreeNAS 9 with ZFS) is a good option for eternal file safety. Is there something like eternal file safety? Would be a great topic to explore in a separate video.

2

I think I can partly answer my question(s) ...

A standard RAID10, you have according to this [Raid10 Setup :: 3 Groups - 3 disks per group] and Survival probability of 100% at 1 disk failure and still 88% with 2 disk failure. The more disks per group, the higher the rate of survival of 2 disk failure scenario. But never 100% for 2 disk failure or even 3 disk failure simultaneously.

I found this interesting article/observation about Raid 5 and 6. In it was a link to this "drives just wear out steadily."

Further, these results validate the Google File System’s central redundancy concept: forget RAID, just replicate the data three times. If I’m an IT architect, the idea that I can spend less money and get higher reliability from simple cluster storage file replication should be very attractive.

Which would mean one should choose RAID1 (array of mirrors, with 3 disk, 2 mirrors). Just add more and more RAID1 volumes (array of mirrors) ... 2 disk failures per RAID array = 100% survivability.

See also this wiki on ZFS, the official FreeNAS HW Recommendations RAID Overview. And this on RAIDZ-3 disk options, #'s and performance/reliability optimisation (Spoiler lots of drives (incl 1 HotSpar)). Interesting too this wiki. Conclusion RaidZ-3? ... if you can afford the amount of disks - $$$. But you can even survive a 3 Disk Failure - with the note that you would not able to read the array/data when rebuilding from a disk failure, I assume. That I can totally live with.

BTW: My basic goal exploring to find the best option of RAID or RAIDZ is to safe data (and occasionally read it) that is mission critical/"eternal file safety." Archiving with occasional reads.

PS: A 5 bay Drobo with only 3 disk (all 4tb) and Dual Redundancy checked = 100% 2 disk failure and the case of volume's directory structure corruption.

A B1200i costs even without drives an gargantuan amount of money, has 2 drive failure survivability but I am not sure about protection against volume's directory structure corruption. And it got only 3x1GbE connections which I guess you can't bundle, and you can't upgrade that damn thing or do anything else with it.

Anybody with more input or a conclusion? Anything else than RAIDZ2 or 3?