Stats are updated only after transaction ends. In case you have a really
long transaction you need something else.
To help myself I made a little Perl utility to parse strace output. It
recognizes read/write calls, extracts file handle, finds the file name
using information in /proc filesystem,
Is there a way to look at the stats tables and tell what is jamming up
your postgres server the most? Other than seeing long running queries
and watch top, atop, iostat, vmstat in separate xterms...I'm wondering
if postgres keeps some stats on what it spends the most time doing or if
there's
