Re: Dynamic Shared Memory stuff

Noah Misch <noah(at)leadboat(dot)com> writes:
> On Tue, Dec 10, 2013 at 07:50:20PM +0200, Heikki Linnakangas wrote:
>> Let's not add more cases like that, if we can avoid it.

> Only if we can avoid it for a modicum of effort and feature compromise.
> You're asking for PostgreSQL to reshape its use of persistent resources so you
> can throw around "killall -9 postgres; rm -rf $PGDATA" without so much as a
> memory leak. That use case, not PostgreSQL, has the defect here.

The larger point is that such a shutdown process has never in the history
of Postgres been successful at removing shared-memory (or semaphore)
resources. I do not feel a need to put a higher recoverability standard
onto the DSM code than what we've had for fifteen years with SysV shmem.

But, by the same token, it shouldn't be any *less* recoverable. In this
context that means that starting a new postmaster should recover the
resources, at least 90% of the time (100% if we still have the old
postmaster lockfile). I think the idea of keeping enough info in the
SysV segment to permit recovery of DSM resources is a good one. Then,
any case where the existing logic is able to free the old SysV segment
is guaranteed to also work for DSM.

regards, tom lane