This one time, at band camp, Paolo Lucente wrote: >Hello Jamie, > >On Fri, Aug 12, 2005 at 04:01:51PM +1000, Jamie Wilkinson wrote: >> Is it true that stamp_inserted and stamp_updated don't get touched unless >> sql_history and sql_history_roundoff are also set? If so, when did this >> change? > >Absolutely true. > >> I was running 0.7.4 for a while and I didn't have these set, recently I >> upgraded and have found data missing. I'm a bit surprised that this might >> be the cause, though the documentation in CONFIG-KEYS leads me to believe >> this is the case... also I wonder what the pmacctd processes might have done >> when trying to write NULL data to a NOT NULL column in the database (namely, >> stamp_inserted). > >This is very likely to be the cause. In previous ( < 0.7.9 pmacct release ) >PostgreSQL schemas, the 'stamp_inserted' column was lacking the trailing >'DEFAULT' modifier into its declaration. This give origin to some troubles >because the 'stamp_inserted' is definitely an optional field: the 'NOT NULL' >was leading to loss of data.
Ok. Does this mean that unless the config options 'sql_history' and 'sql_history_roundoff' exist, then pmacctd will not write time stamps to the database? >Please, alter the table manually (using the 'ALTER TABLE ... SET DEFAULT') >or apply the SQL script (of the new pmacct release) to create the new table of >the same version you are currently using. I've done so, but I've also added these two config options back to my config file, and I'm seeing a lot of 'We are missing data' errors in the syslog.
