Re: [GENERAL] Statistics mismatch between n_live_tup and actual row count

It seems that we are currently running 8.4.3 on the server we are encountering the problem. Will upgrade to 8.4.9 and then will come back with a test case if we still see the issue Thanks again for your help. -- View this message in context: http://postgresql.1045698.n5.nabble.com/Statistics-

Re: [GENERAL] Statistics mismatch between n_live_tup and actual row count

Thanks for the reply Tom, will try and construct test case. Have been unable to replicate in a simple test the sort of updates that the table out in the wild is seeing, so may impact that issue of vacuum finding unrepresentative sample, maybe. Will try harder! When you say recent releases, does

Re: [GENERAL] Statistics mismatch between n_live_tup and actual row count

86% of the updates are HOT updates! The difference between the first and second line of this image above is that 366 updates happened of which 299 where HOT. And a vacuum on the table was run. Look at the retuples number c

Re: [GENERAL] Statistics mismatch between n_live_tup and actual row count

I am also seeing a drift in the n_live_tup value compared to actual row count on the table on PG9.0.6 It drifts after a vacuum , you can bring it back closer to the actual number by running ANALYSE several times, you can lock it back into the right value with a vacuum full, but then if you run a v