On Tue, Aug 30, 2005 at 06:07:24PM +0200, Michael Fuhr wrote: > tables, and a VACUUM might start or complete immediately after you > issue the query but before you read the results). This method is > therefore unreliable. I intend to do the VACUUM FULL during quiet hours, thus the chance of fitting exactly to the time that VACUUM started and it is not reflected in the tables is quite small. And even if it would happend, very likely it will affect only one user, who may get around hitting "refresh" button. > > > What problem are you trying to solve? If we knew what you're really > trying to do then we might be able to make suggestions. I have database, which gets around 240 000 new lines each day and about the same is also deleted each day. The table has something around 8M lines in average and simple query takes about 70s to complete(V210 1x UltraSPARC-IIIi). As this time is quite high, I need "defragment" database on daily basis. These queries get visualized in web application. My problem is, how to make the web application aware that maintainace (VACUUM FULL) is in place, but the database is not down. I really would not like to do it via extra status table, while sometimes it may happend, that someone will run VACUUM FULL ad-hoc-ly in good-faith and will forget to update the status table. ---------------------------(end of broadcast)--------------------------- TIP 5: don't forget to increase your free space map settings
