[EMAIL PROTECTED] wrote:

I see, so replicated volumes are static, rather than constantly catching up?

Currently, AFS support read-only replication but not read-write replication.

A .readonly volume is paired to a read-write volume. The "vos release" command will copy the current contents of the read-write volume to the .readonly volume. A .readonly volume may have replicas on multiple file servers.

4) If the read-write server fails, how does OpenAFS handle failing over to
the replicated backup? When the original master comes back up, how
transparently / gracefully does this happen?


For read-write, you can't see it. For readonly, a volume is a volume.

So, mounting a read-write volume doesn't transparently fail and promote the backup.

There are two types of mount points in AFS: Normal and read-write. A normal mount point will always prefer a readonly volume over a read-write volume. A read-write mount point will always use the read-write volume. Once a read-write mount point has been traversed in a path, all subsequent mount points will be evaluated as read-write mount points.

Clients will failover for .readonly volumes from file server to file server. There is no failover for read-write volumes because there is only a single instance of a read-write volume.

Jeffrey Altman

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply via email to