Sebastien Guilbaud wrote:
I always have a difference of one hour in my database
between these columns :
stamp_inserted stamp_updated
2006-02-02 10:10:00 2006-02-02 09:16:26
I've tried to activate debugmode, and INSERT's contain
timestamps with one hour variation :
INSERT INTO acct_v2 (stamp_updated, stamp_inserted, ip_src, ip_dst, port_src, port_dst, ip_proto, mac_src, mac_dst,
packets, bytes) VALUES (ABSTIME(1138868666)::Timestamp, ABSTIME(1138872000)::Timestamp, '0.0.0.0', 'x.x.x.x', 0, 0, 0,
'0:0:0:0:0:0', '0:0:0:0:0:0', 2, 96)
I don't know whether this comes from nfacct or from a postgres
setting :-(( Is anybody using postgresql 8.1 to store nfacct
data ??
I ran into some similiar symptoms when I was using sql_history to create
hourly counters without the roundoff parameter - counters had
stamp_inserted and stamp_updated values in different hours rather then
in the same hour. Adding sql_history_roundoff fixed that problem...
sql_history: 1h
sql_history_roundoff: h
Wim
