On 11/9/09, Jon Burgess <[email protected]> wrote: > On Mon, 2009-11-09 at 21:17 +0100, Peter Körner wrote: >> > >> > The changes in changeset 3073502 are completely missing from the >> > minute-replicate diffs. They appear in both the minute and hourly >> diffs. >> > >> >> I checked node 559977476 [1] that has been created in changeset >> 3073502 >> [2]. I can't find them in the minute-replicate diffs. They should be >> in >> the index between 000/084/536 - 000/084/537 but grepping for this id >> returns nothing. > > I looked on the mapnik database being fed by these diffs and it does not > show this node either. There is not a single reference to changeset > 3073502 in any of the 100 - 999 diffs in the 000/084 directory. > > References to the changesets on either side (3073501 & 3073503) appear > in the 539.osc.gz file. I have no idea why this one has gone missing.
possibly it's a coincidence, but changeset 3073502 was committed in transaction 174978321, which is the txnMax of replication diff 84538. i don't see anything special about that number, though. is it possible that the connection pooling in the apache DBCP BasicDataSource means that the transaction state query is occurring in a connection which has txn 174978321 committed, but the replication query happens in one where it hasn't been committed yet? cheers, matt _______________________________________________ dev mailing list [email protected] http://lists.openstreetmap.org/listinfo/dev
