Norbert Hoffmann <[EMAIL PROTECTED]> writes: > spaetz wrote: > >>On Tue, Aug 19, 2008 at 08:27:34PM +0200, Norbert Hoffmann wrote: >>> |08-08-19-11:24:25 INFO Reset 1 tileset (12,1238,1515) by user pavel (uuid: >>> 2142848450). Cause: NoData >> >>This is very weird and means that python really jumps by 2 hours. Never seen >>that before.... > > It's not so seldom and started IIRC even before tahngo. If you look at > http://tah.openstreetmap.org/Request/show/ now and 1 minute later again, > you'll see some requests where the "last client contact" stays at "0 > minutes". Those requests seem stored with a time in the future (up to those > 2 hours, i think). > > Those jumps happen for some hours to days and then are gone for a long > time.
You can see them very well at http://munin.openstreetmap.org/openstreetmap/tah.openstreetmap-tah_processed.html I believe all those dips in the graph happen when Munin is catching tah when it is 2 hours ahead of time. Matthias _______________________________________________ Tilesathome mailing list [email protected] http://lists.openstreetmap.org/listinfo/tilesathome
