Manoj Kumar <manojkumar@ameyo.com> writes:
> Adding more data to support my point in form of attachment.
Well, those t-something files are certainly temporary tables,
so what you've got there is that some session has created a bunch
of quite large temporary tables (the "t20_" files) and another has
created a bunch of rather smaller ones (the "t62_" group).
However, this is not direct evidence of PG misbehavior. You've
offered no reason to think that the sessions that made those files
have exited. Even if you're sure they have, we can't do anything
to help you on the basis of the zero facts you've supplied so far
about what the sessions were doing.
If you can provide a reproducible test case that leaks temp tables,
we could get to the bottom of it pretty quickly, I'm sure. But
we cannot help you on the basis of what you've shown us.
regards, tom lane
