help? trouble setting Shared Memory parameters in OSX kernel (fwd) - Mailing list pgsql-hackers

From R Blake
Subject help? trouble setting Shared Memory parameters in OSX kernel (fwd)
Date
Msg-id 2147483647.1047836773@[172.30.10.6]
Whole thread Raw
List pgsql-hackers
hi all,

tried this on pgsql-admin & darwin-kernel lists, and unfortunately, not a 
'nibble' ...

can ne1 here spare a moment?

thanks!

richard

---------- Forwarded Message ----------

hi all,

i've successfully built postgreSQL 7.3.2-STABLE on Mac OSX 10.2.4.

much trouble launching it led me to the discussions
(<http://www.postgresql.org/docs/view.php?version=7.3&idoc=1&file=kernel-re
sources.html>) on setting shared memory parameters -- especially SHMMAX --
in OSX's "/System/Library/StartupItems/SystemTuning/SystemTuning"

so, as instructed, i changed the defaults (near the bottom of SystemTuning):
   sysctl -w kern.sysv.shmmax=4194304   sysctl -w kern.sysv.shmmin=1   sysctl -w kern.sysv.shmmni=32   sysctl -w
kern.sysv.shmseg=8  sysctl -w kern.sysv.shmall=1024
 

to:
   sysctl -w kern.sysv.shmmax=33554432 # bytes   sysctl -w kern.sysv.shmmin=1   sysctl -w kern.sysv.shmmni=32   sysctl
-wkern.sysv.shmseg=8   sysctl -w kern.sysv.shmall=8192 # 4K pages
 

and defined (in postgresql.conf):
       shared_buffers = 2048       max_coinnection = 1024

and rebooted.

unfortunately, immediately after system launch, i still couldn't launch
postgres

eventually, a random check with "sysctl -a" gave:
       kern.sysv.shmmax: -1       kern.sysv.shmmin: -1       kern.sysv.shmmni: -1       kern.sysv.shmseg: -1
kern.sysv.shmall:-1
 

OBVIOUSLY, a problem!

i CAN enter the sysctl settings at the command line, verify success with
'sysctl -a', and then postgres launches/operates just fine .....

QUESTIONS:

any idea why these setting are being written at boot to values of "-1"?

how do i get these setting to 'stick'?

thanks,

richard





pgsql-hackers by date:

Previous
From: Hannu Krosing
Date:
Subject: Re: Roadmap for FE/BE protocol redesign
Next
From: Teodor Sigaev
Date:
Subject: Current CVS compile problem