[Gluster-devel] AFR write completion? AFR read redundancy?

[Anand Avati]

>
> Will self healing prevent an inconsistent cluster from
> happening?  I.E. Two node cluster, A+B.
>
> 1) Node A goes down
> 2) Write occurs on Node B
> 3) Node B goes down (cluster is down)
> 4) Node A comes up -> cluster is inconsistent since B
> is not yet available.  Cluster should still be "down".


This is not assured to work. The intersection of the two subsets of
subvolumes before and after a group (subset) of nodes are added or removed
should not be empty.

> Any plans for such a feature?  HA is great, but so
> > is error resilience/correction. :)
> >
> > we have thoughs of checksum based error
> > detection/correction. Infact one of
> > the users has submitted a n ECC based translator
> > (yet to be reviewed seriously). But no plans of
> > making it part of AFR.
>
>
> So this ECC AFR would work on a single subvolume then?


ECC (not ECC AFR) would have a signature subvolume which would have
redundancy bytes capable of detecting/correcting errors in the 'main'
subvolume. This has nothing to do with AFR.

> We are in fact even thinking of raid5 or raid6
>
> > kind of translator which does checksum based error
> > detection/recover to some extent.
>
>
> Would this be a part of AFR, or a completely new
> translator?


completely new.

thanks,
avati