Re: [PERFORM] in-transaction insert performance in 7.5devel

2004-06-17 Thread Merlin Moncure
> Was that 15% before or after updating from CVS? > > The more I think about the looping aspect the less I like it, so I'd > prefer not to pursue making the unlock change for real. But if it's > really a 15% win then maybe we need to... > > regards, tom lane After. So far

Re: [PERFORM] in-transaction insert performance in 7.5devel

2004-06-11 Thread Tom Lane
"Merlin Moncure" <[EMAIL PROTECTED]> writes: > Tom Lane wrote: >> Does it help if you change FlushBuffer to release buffer lock while >> flushing xlog? > Putting your change in resulted in about a 15% increase in insert > performance. There may be some quirky things going on here with NTFS... >

Re: [PERFORM] in-transaction insert performance in 7.5devel

2004-06-11 Thread Merlin Moncure
Tom Lane wrote: > "Merlin Moncure" <[EMAIL PROTECTED]> writes: > > I am batch inserting insert statements into a database with fsync = on. > > My single disk system is on a 10k drive...even though I am inside a > > transaction there is at least 1 file sync per row insert. > > Are you certain you'

Re: [PERFORM] in-transaction insert performance in 7.5devel

2004-06-11 Thread Tom Lane
"Merlin Moncure" <[EMAIL PROTECTED]> writes: > I am batch inserting insert statements into a database with fsync = on. > My single disk system is on a 10k drive...even though I am inside a > transaction there is at least 1 file sync per row insert. Are you certain you're inside a transaction? T

Re: [PERFORM] in-transaction insert performance in 7.5devel

2004-06-11 Thread Rod Taylor
> As I understand it, sync() is never called anymore. mdsync() hits the > all the files 1 by 1 with an fsync. My understanding of the commit > process is that 30 tps is quite reasonable for my hardware. Sorry. I didn't see the version in the subject and assumed 7.4 on a Linux machine with exce

Re: [PERFORM] in-transaction insert performance in 7.5devel

2004-06-11 Thread Merlin Moncure
Rod Taylor wrote: > On Fri, 2004-06-11 at 14:40, Merlin Moncure wrote: > > I am batch inserting insert statements into a database with fsync = on. > > My single disk system is on a 10k drive...even though I am inside a > > transaction there is at least 1 file sync per row insert. > > Which filesys

Re: [PERFORM] in-transaction insert performance in 7.5devel

2004-06-11 Thread Rod Taylor
On Fri, 2004-06-11 at 14:40, Merlin Moncure wrote: > I am batch inserting insert statements into a database with fsync = on. > My single disk system is on a 10k drive...even though I am inside a > transaction there is at least 1 file sync per row insert. Which filesystem? PostgreSQL isn't issui

[PERFORM] in-transaction insert performance in 7.5devel

2004-06-11 Thread Merlin Moncure
I am batch inserting insert statements into a database with fsync = on. My single disk system is on a 10k drive...even though I am inside a transaction there is at least 1 file sync per row insert. Is this normal? On my hardware, which is pretty quick, I am limited to about 50 inserts/sec. If