So why do we nerds use NAS/Server storage with RAID systems? Because we know hardware can fail, especially mechanical hardware.
And there are different variants of RAID (redundant array of independent disks) systems
- RAID 0 (All disks just appear as one)
- RAID 1 (One disk is mirrored to another one)
- RAID 10 (Half of the disks is mirrored to the other half)
- RAID 5 (Redundant parity information is stored so that 1 out of N disks can fail without data loss)
- RAID 6 (Same as RAID 5 with one additional spare disk, so that 2 out of N can fail)
Back to topic, why am I writing ‘FUCK YOU, probability’?
The idea for RAID 5/6 came out of the typical probability how often a mechanical disk may fail. So if you’re almost paranoid you’re choosing RAID 6 for your Data to be ‘absolutely’ sure you won’t loose data.
Yeah, of course, until you come to that point, like me, when over a fucking weekend 3 out of your 5 disk RAID 6 system fail.
Yes, you’ve read correct. Three disks of my system failed, what means that my data is almost gone.
In the hours I tried to force the RAID array again into a working state I encountered that 2 out of the 3 failed disks are worth a brick. But the last was still responding, a lot of ATA errors but still responding.
Good, I put that RAID together again and am currently trying to rescue data with ddrescue and other tools. (Already running for 3 days… yeah, USB2.0 and 6TB of data… NARF!)