On Mon, May 16, 2011 at 03:05:20PM +0100, Guido Trotter wrote: > On Tue, May 3, 2011 at 5:05 PM, Adeodato Simo <[email protected]> wrote: > > Hi, > > > > > On Tue, May 3, 2011 at 17:59 +0200, Iustin Pop <[email protected]> wrote: > >> On Tue, May 03, 2011 at 03:22:10PM +0100, Adeodato Simo wrote: > >> > This commit prepares the call to _VerifyFiles for the case when the > >> > master > >> > node is not one of the nodes that's being verified (which will be the > >> > case > >> > for all node groups but one). We fix it by always passing master info and > >> > checksums to _VerifyFiles, which ensures there's a cluster-wide > >> > consistency > >> > check. > > > >> The master node can be a non-vm_capable node, so I'm not sure if > >> choosing it unconditionally is the best thing… > > > > AFAIK, verifying against master's checksums was the previous behavior, > > and the one that the reworked _VerifyFiles from aef59ae preserved. This > > commit is just working with that... > > > > What do we do with this one? Should we have a separate patch that > choses "a vm_capable node outside the group" rather than choosing > always the master?
My reading of Dato's explanation is that the current code is right, and that we pick the first vm_capable node. If not true, then we need to do something like that. iustin
