checksum-xattr.diff [CVS update: rsync/patches]

[foner-rsync at media.mit.edu]

    Date: Mon, 2 Jul 2007 21:18:57 -0400
    From: "Matt McCutchen" <hashproduct+rsync at gmail.com>

    The technique Wayne and I are discussing assumes only that the clock
    on *each side* never steps backwards.

Um, and note, btw, that the pathological FSB-spread-spectrum/NTP
interaction I mentioned in my first message was causing a whole
-bunch- of backwards steps, over several months, until it was noticed.
I don't recall their magnitude, but I think it was a backwards step of
at least a second every few tens of minutes, until after quite some
time NTP simply exceeded its tolerance and punted, whereupon the clock
ran away.  But since it was -almost- holding it together, for days or
weeks at a time...

And the machine was an NFS server.  So in fact this scheme would have
been leading to a whole bunch of sporadic "why was this cache
inaccurate?" failures for a long time if rsync had been using this
strategy and someone had been using it against that server.