On 2/14/2011 7:12 PM, Aleksey Tsalolikhin wrote:
> I then shut down PostgreSQL and then started it again, and usage dropped
> to 6%.

Since a successful TRUNCATE does actually truncate(2) the heap and 
reinitializes all related files (indexes, toast heap and index), the 
disk space could not have been held hostage by anything related to that 
table itself.

Is it possible that there was some stale connection that held temp 
files? PostgreSQL keeps temp files like sort sets in 
$PGDATA/base/pgsql_tmp and temp tables inside the database directory. 
They do normally go away when they are no longer needed. Sort sets when 
the statement ends, temp tables when the session ends. Since a 
postmaster restart did free all that space, I suspect a temp table.

You mentioned using some Bucardo tool to verify things. Does that tool 
eventually create temp tables and could it be, that it left a stale 
connection behind keeping those temp tables around?


Jan

-- 
Anyone who trades liberty for security deserves neither
liberty nor security. -- Benjamin Franklin