Re: All Production releases of PostgreSQL hung because of
Reini Urban <[email protected]>
| Newsgroups | gmane.comp.db.postgresql.cygwin,gmane.os.cygwin |
|---|---|
| Message-ID | <[email protected]> |
Jean-Pierre Pelletier schrieb:
> There is currently no production release of PostgreSQL that runs more
> than 5 concurrent sessions
> in Windows because the releases available for Cygwin, 7.4.3 to 7.4.5 are
> built with Cygserver
> which has a bug that have PostgreSQL hung with more than 5 concurrent
> connections.
Hmm, that would explain the crash at the parallel_schedule regression
suite. ipcs showed 5 semaphores, and one shm, when postmaster crashed.
serial tests do work fine. max_connections=100 also works fine.
Do you have any link where this is discussed/analyzed?
> I am planning to ;revert to PostgreSQL 7.1 which is the latest cygwin
> package built with cygipc (instead of cygserver).
> ;
> It would be nice to have the latest releases of PostgreSQL rebuilt with
> cygipc and available as cygwin packages.
No problem.
If it's really a buggy cygserver I'll do a interim cygipc version.
I already had one, but it clashed with cygserver.
But I really want to try to debug cygserver then.
--
Reini Urban
http://xarch.tu-graz.ac.at/home/rurban/
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match