Vaccuming dead rows on busy databases

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


Is there a general strategy out there for removing dead rows on
busy databases? I've learned that both a vacuum and a vacuum full
will not remove dead rows if there is an open transaction anywhere
on the entire cluster. Closing open transactions is not always a
viable solution, especially for periodic cron-driven vacuums. What
do people generally do to get around this problem? I can think of
a few possible solutions, but wanted to see if there was a best
practice before I went too much further. All my solutions are
external as well: I'd like to see something like VACUUM FULL WAIT. :)

- --
Greg Sabino Mullane greg@turnstep.com
End Point Corporation
PGP Key: 0x14964AC8 200612181331
http://biglumber.com/x/web?pk=2529DF6AB8F79407E94445B4BC9B906714964AC8

-----BEGIN PGP SIGNATURE-----

iD8DBQFFht6qvJuQZxSWSsgRArZDAKD9zfrsuysep38o+UaW/2tDL/AdvACguZIO
jd0JERBi0AwBBZj+xu80Qtg=
=vSuy
-----END PGP SIGNATURE-----