[Openstack] [Swift] Erasure code durability and overhead in small clusters
Mark Kirkwood
mark.kirkwood at catalyst.net.nz
Wed Mar 14 00:33:31 UTC 2018
On 14/03/18 11:47, Clay Gerrard wrote:
>
>
> On Tue, Mar 13, 2018 at 3:05 PM, Mark Kirkwood
> <mark.kirkwood at catalyst.net.nz <mailto:mark.kirkwood at catalyst.net.nz>>
> wrote:
>
> To me this suggests that a certain minimum number of *hosts* per
> region is needed for a given EC policy to be durable in the advent
> of host outage (or destruction). Is this correct - or have a
> flubbed the calculations?
>
>
> Looks correct to me. Ideally you'd have more nodes per region than
> k+m; otherwise a reboot would mean WAN traffic - or if that's not
> possible - temporary unavailability.
>
>
Awesome thanks!
Doing a bit more playing about leads me to think that for a *single
region* EC policy we can get a tighter lower bound on the number of
hosts: I'm calculating it as (k+m)/m. So in the case I was playing with
before (k=8, m=2), then I need 5 hosts in my region. While testing does
seem to support this, does it sound reasonable?
Cheers
Mark
More information about the Openstack
mailing list