Re: kern.bufcachepercent

2009-11-04 Thread Luis Useche
Maurice: Thanks for pointing that out. Bob: At this point this is probably irrelevant. In any case, I found it in the officiel webpage http://www.openbsd.org/plus46.html. Thanks for your help! Luis On Wed, Nov 4, 2009 at 10:42 AM, Bob Beck wrote: > I don't know what version of plus46.html you

Re: kern.bufcachepercent

2009-11-04 Thread Bob Beck
I don't know what version of plus46.html you are looking at - but that text doesnt' appear in any version I look at. Of course it is in the cvs commit log, but that's not the same thing. That same commit was backed out before 4.6 - and has since gone back into current. 2009/11/4 Luis Useche : > O

Re: kern.bufcachepercent

2009-11-04 Thread Maurice Janssen
On Wed, Nov 04, 2009 at 10:26:50AM -0500, Luis Useche wrote: >OK. Sorry for the noise. In any case, this change is in the 4.6 >changelog (twice, http://www.openbsd.org/plus46.html): > >"Added dynamic buffer cache sizing. The sysctl kern.bufcachepercent >will allow you to specify a high-water mark a

Re: kern.bufcachepercent

2009-11-04 Thread Luis Useche
On Tue, Nov 3, 2009 at 11:44 PM, Bob Beck wrote: > 2009/11/3 Luis Useche : > >> >> I read in the 4.6 changelog that his was part of the release. >> >> Am I missing something? Do I have to recompile? Or this is just a bug? > > Yeah you are missing something. Listen to the *whole* presentation and >

Re: kern.bufcachepercent

2009-11-03 Thread Bob Beck
2009/11/3 Luis Useche : > > I read in the 4.6 changelog that his was part of the release. > > Am I missing something? Do I have to recompile? Or this is just a bug? Yeah you are missing something. Listen to the *whole* presentation and read the *whole* changelog. This is *not* in 4.6 It is in cu

Re: kern.bufcachepercent

2009-11-03 Thread Robert
On Tue, 3 Nov 2009 19:34:24 -0500 Luis Useche wrote: > Hi Guys, > > Today I watched the presentation by Bob Beck about the OpenBSD VFS > hacking where he talked about the buffer cache and how it recently > changed (among many other things). Following his advice, I tried to > increase my buffer c