On Wed, Jan 22, 2014 at 10:17 AM, Noah Misch <noah@leadboat.com> wrote:
> Yeah, abandoning the state file is looking attractive.
Here's a draft patch getting rid of the state file. This should
address concerns raised by Heikki and Fujii Masao and echoed by Tom
that dynamic shared memory behaves differently than the main shared
memory segment. The control segment ID is stored in the System V
shared memory block, so we're guaranteed that when using either System
V or POSIX shared memory we'll always latch onto the control segment
that matches up with the main shared memory segment we latched on to.
Cleanup for the file-mmap and Windows methods doesn't need to change,
because the former always just means clearing out $PGDATA/pg_dynshmem,
and the latter is done automatically by the operating system.
Comments?
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company