| |||||||||||||
PostgreSQL, I wishPostgreSQL, I wishPosted Aug 22, 2012 18:56 UTC (Wed) by raven667 (subscriber, #5198)In reply to: PostgreSQL, I wish by dlang Parent article: MariaDB: Disappearing test cases or did another part of MySQL just become closed source?
What's your opinion on pgtune fitting this bill? There seems to be only a handful of truly critical postgresql performance tuning knobs with quickly diminishing returns that will be highly workload-specific and not hardware-specific. I wonder if there is any idea in the postgresql community to just have it detect cpu/memory at startup time and automatically adjust the internal buffers unless overridden by the config file, making pgtune unnecessary.
(Log in to post comments)
PostgreSQL, I wish Posted Aug 22, 2012 19:31 UTC (Wed) by andresfreund (subscriber, #69562) [Link]
Some of that is happening (wal_buffers is autotuned since 9.1, max_fsm_* doesn't exist anymore since 8.4, default_statistics_target has a sane default since 8.4, 9.3 won't need adjustment of sysv /proc settings for most things), in most of the other situations its either hard to do or it very much depends on the environment. Many pg instances don't run on a dedicated machine. Many database servers serve multiple pg clusters at the same time. Many database servers run in virtualized environments where you get better average performance if you use less memory.
There are other variables where nobody fought enough to get them changed. Other changes are beneficial in a wide range of systems but hurt others badly...
|
|||||||||||||