Re: [HACKERS] TODO: Determine optimal fdatasync/fsync, O_SYNC/O_DSYNC options
Joshua D. Drake [EMAIL PROTECTED] writes: Doesn't this exist in: src/tools/fsync? No, that only tells you if the various alternatives *work*. Not which one performs best. regards, tom lane ---(end of broadcast)--- TIP 1: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
Re: [HACKERS] TODO: Determine optimal fdatasync/fsync, O_SYNC/O_DSYNC options
Tom, JD, No, that only tells you if the various alternatives *work*. Not which one performs best. Performs best is pretty elusive. On Solaris 10, for example, fdatasync does best on pgbench on x86, but opendatasync does best on DBT2 on Sparc. I'm not certain we can really determine this with any accuracy. -- --Josh Josh Berkus PostgreSQL @ Sun San Francisco ---(end of broadcast)--- TIP 4: Have you searched our list archives? http://archives.postgresql.org
Re: [HACKERS] TODO: Determine optimal fdatasync/fsync, O_SYNC/O_DSYNC options
Josh Berkus josh@agliodbs.com writes: Performs best is pretty elusive. On Solaris 10, for example, fdatasync does best on pgbench on x86, but opendatasync does best on DBT2 on Sparc. I'm not certain we can really determine this with any accuracy. Yeah, if it were easy we've have done it already :-(. One thing that would be useful though is to try to gather datapoints like the above for a range of different operating systems. Even if we could tell people here are the two or three settings worth trying for your OS, we'd be ahead of the game. regards, tom lane ---(end of broadcast)--- TIP 9: In versions below 8.0, the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match