On Sun, 5 Oct 2014, Patrik Horník wrote:

I want to ask for best solution for replacing device in RAID 6 array.

- We discussed couple of years ago hot replacing the drive. I see it
is now supported in mdadm. Is it rock stable and can be used on
production data? I have kernel and mdadm from debian testing, 3.14-2
and 3.3-2. Were there some important fixes from these versions? I have
this server on UPS but not on aggregator so there is real danger of
power outage and need to shutdown the array during operation.

I have successfully used the replace function without ill effects, and I 
believe this was with a 3.8.x kernel. The function itself has been around 
since 3.3 or so.

I don't want to give you any recommendations, but next time I need to 
replace a failing drive that is still operational enough to stay in the 
raid, personally I will use the replace function. The code is a few years 
old by now and I imagine it has been used quite a lot by now.

-- 
Mikael Abrahamsson    email: swmike@swm.pp.se