On Oct 7, 2011, at 3:48 PM, René Fournier wrote:
> Well, I changed them at the shell and in sysctl, and restarted (on both machines).
Check them at the shell. If they're not what you set in sysctl.conf, then your sysctl.conf is not correct, and since
youcan't change them after startup, you won't be able to launch pg with anything but a tiny shared buffers setting.
Thereare restrictions on them, such as shmall being an even multiple of the VM page size, and some relationship or
otherbetween shmall & shmmax but I don't remember the details.
> So one thing that's odd, I have two version of postgresql apparently running (although I've removed 9.0 from
Macports)…Not sure if this is the cause. Any other suggestions?
Yes. Figure out where the logging for the postgres instances you are trying to launch are going, and read the log
messages.Or maybe they're just going to console--check that first.
You'll have a bit of difficulty getting help here, because you're not likely to find people who know what those
wrappers(daemondo & postgresql84-server.wrapper) actually do, since they're not part of postgresql. One thing you
shouldknow, it is possible to have 2 versions of pg running at the same time, but it is not possible to have 2 versions
runningon the default port at the same time, or in the same data directory ("cluster" in pg terminology).
Also, you don't have any actual pg instances running there. What you have is some kind of wrapper that tries to launch
pg.That wrapper is failing to launch, and either looping, or quitting and being relaunched--depending on how it is set
up,which I have no idea about.
--
Scott Ribe
scott_ribe@elevated-dev.com
http://www.elevated-dev.com/
(303) 722-0567 voice