On Wed, Sep 8, 2010 at 10:04 AM, Rob Austein <s...@isc.org> wrote: > I don't see any locking strategy (either modifying rsync or creating a > new RPKI object to represent a lock) as likely to work. I can go into > details if necessary, but in short there are just too many different > ways that a relying party could end up with a potentially inconsistent > collection of objects, and this problem goes beyond a single directory > ("publication point"), or the rsync protocol. > > I see no alternative but to place the burden of assembling a > consistent view on the relying party. This may mean that the relying > party's view lags the latest published data in some cases, and that a > relying party that's just starting up (has no cache of valid results > from previous retrieval runs) may end up with an incomplete picture > portions of the universe. So be it. > > Welcome to the world of loosely consistent databases.
as an aside, rsync issues can mostly be avoided with: morr...@donkey:~$ rsync --help | grep delay --delete-delay find deletions during, delete after --delay-updates put all updated files into place at transfer's end or that sort of thing (--delay-updates) is a common thing to use in syncing large sets of data where other parties are reliant on a consistent view (and changing that view under them is considered rude). For this discussion though it strikes me that there could be lots of reasons that an RP sees a slightly different view of the publication point over time, different from what another RP sees, or the publisher intends inside a short period of time at least. Checking what you download with what is supposed to have been there (manifest checks) are still required, or prudent at least. -chris > _______________________________________________ > sidr mailing list > sidr@ietf.org > https://www.ietf.org/mailman/listinfo/sidr > _______________________________________________ sidr mailing list sidr@ietf.org https://www.ietf.org/mailman/listinfo/sidr