On Thu, 2008-01-10 at 07:52 +0100, Gábor Farkas wrote:
> the remaining 3 were only idle-in-transaction at that point. so if i
> would keep checking for idle-in-transaction processes, the list of them
> would keep changing.
>
> are you saying, that a process should NEVER be idle-in-transaction? not
> even for a short time? (like some seconds?)
It's OK to be idle-in-transaction, but not OK for that state to last for
days.
> also, even if it is wrong, can an 'idle-in-transaction' connection that
> was opened today block the vacuuming of rows that were deleted yesterday?
Yes, if the rows were deleted after the connection started.
--
Simon Riggs
2ndQuadrant http://www.2ndQuadrant.com