jdd@dodin.org wrote:
Le 26/10/2017 à 19:27, Dave Howorth a écrit :
On Thu, 26 Oct 2017 18:37:59 +0200 "jdd@dodin.org" <jdd@dodin.org> wrote:
Le 26/10/2017 à 18:16, Wols Lists a écrit :
Hard drives at the bottom, raid on that, lvm on that, and probably btrfs on that.
I hope you wont ever have a faulty drive...
That's the main reason for having that setup. When a drive fails the machine carries on whilst you buy a replacement and swap it for the failed drive, which rebuilds.
not always, not always like you want, not always when you want.
I have to say - I have been running some 20 systems (hosted externally) with RAID1 for more than 12 years. For the last 5-6 years, 1 or 2 drives fail per annum, sometimes more. They are replaced on demand, they rebuild and all done. Exactly as Dave describes it. Internally we use hot spares on every system, no complaints there either.
this setup is extremely expensive and is interesting is you want a system as much as possible always running.
Not _extremely_ expensive, only the price of a second drive. Apart from high availability requirements, it's very useful if you want to avoid restoring a backup just because of a disk failure. The small price of the extra drive is easily paid back.
there are many situation where rebuilding a raid array can be very time consuming and error prone
Yes, the bigger the array, the longer it takes and that does become a problem. Error prone? Not in my experience.
for example, do you have at hand an exact replacement for any disk in the raid array. What if one fail and you have to wait three days to get an other?
You don't need an exact replacement - only one that is the same size of bigger. -- Per Jessen, Zürich (12.9°C) http://www.hostsuisse.com/ - virtual servers, made in Switzerland. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org