Hi,
I'm trying to understand object storage replication between two ceph
clusters in two zones in a single region. Setting up replication itself
isn't the issue, it's how to ensure high availability and data safety
between the clusters when failing over.
The simplest case is flipping the primary cluster between the two zones,
just to test that we *can* do that. We expect that there is data that
was written to the original primary that has not be replicated to the
backup cluster yet, just because people are constantly writing to it. In
that case, is there any way to ensure that data that was written to the
original primary, but not replicated yet to the other zone, will be
copied? Or is that data only in one place until we flip back over to the
original primary?
The same situation would happen upon flipping primary responsibilities
back to the original, where there would be data that had just been
written that hadn't been replicated yet. Now do we have data on both
sides that exists only in one place with no way to ensure that it is
made consistent between the two zones?
Sorry if that was confusing, but it's hard to describe without seeing
the gestures I'm making as I'm explaining it :)
bab
--
Brian Button
bbut...@agilestl.com | @brianbuttonxp | 636.399.3146
http://www.agileprogrammer.com
_______________________________________________
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
