Alvaro Herrera <alvherre@commandprompt.com> writes:
> Tom Lane wrote:
>> Another point to keep in mind, if you are trying to analyze files
>> belonging to a live database, is that what you can see in the filesystem
>> may not be the "current" contents of every page. For typical access
>> patterns it'd be unsurprising for the visible index pages to lag behind
>> those of the heap, since they'd be "hotter" and tend to stay in shared
>> buffers longer.
> Hmm, I think the files come from a PITR slave that's not online. I'll
> ask to be sure.
8.1 didn't have restartpoint code, and of course bgwriter isn't running;
so I believe filesystem pages on a PITR slave could be arbitrarily far
out of date if the corresponding shared buffer got touched regularly.
Try doing the analysis on the master immediately after a CHECKPOINT
command.
regards, tom lane