Marco Colombo wrote:
> Well, many defaults are hardcoded into a file now. I'd like to see
> 'auto' among possible values of parameters, e.g.:
> with PG wild guessing reasonable values based on system specs. It may
> be a awful piece of code (getting system info is very platform
> specific), and sometimes the guess may be wrong.
The problem here is that the amount of shared memory a system can
allocate is hard to discover any other way than starting the server and
seeing if it works. So doing what you advise will leave the database
unable to start on any system that hasn't gotten the right OS kernel
tweaks done first. If there was an obvious 'auto' setting, I guarantee
you newcomers would use it without consulting the documentation first,
then tell everyone it doesn't work because their server broke when it
was enabled. There is a technique used when you run initdb to try and
detect the shared memory allocation range possible, but it really
doesn't scale very well into the sizes needed to be useful for this.
--
Greg Smith 2ndQuadrant US Baltimore, MD
PostgreSQL Training, Services and Support
greg@2ndQuadrant.com www.2ndQuadrant.us