That's not a standard feature of Raid 5. The data on raid 5 is spread across the disks n-1 parts data, 1 part parity. When you add another drive you have to recalculate this parity.
Shouldn't that happen automatically though, in the background?
I upgraded my 4x 1TB array to 4x 2TB.
Steps:
- Swap out the drives one at a time. I might have had to run some Initialize tool in the software, I don't remember that part anymore, but recalculating the parity was done automatically, though it does take a few hours.
- Use Acronis Disk Director to resize the existing partitions to use the newly-available space.
It was fairly painless.
The array
is vulnerable while it's redoing the parity though, so that's the riskiest part of it.
The other risk is those non-recoverable read errors. Drive sizes have been getting bigger, but the read error rate hasn't been keeping pace.
Interesting thing, as I look again at the specsheet:
Normal drive specs say "<1 in 10^14" as the non-recoverable error rate for a consumer-grade drive.
But the specsheet for the enterprise-grade drive says "<
10 in 10^16."
Huh.
So...I'm not good with prob&stat.
How does
10-in-10^16
compare to
1-in-1^14
?
Or did they make a typo in the datasheet, and that 10 should be a 1?