Date: Thu, 12 Jul 2001 14:59:03 -0700 (PDT) From: Matt Dillon <dillon@earth.backplane.com> To: "Juha Saarinen" <juha@saarinen.org> Cc: "'Jonathan Fortin'" <jfortin@akalink.com>, <freebsd-stable@FreeBSD.ORG> Subject: Re: RE: Regarding New FreeBSD BenchMark From Sysadmin Mag (left out a fiew tuning options) Message-ID: <200107122159.f6CLx3M51540@earth.backplane.com> References: <004b01c10b1c$ddb040b0$0a01a8c0@den2>
next in thread | previous in thread | raw e-mail | index | archive | help
::: hw.ata.wc = 1 : :How do you enable write-caching on SCSI drives? The camcontrol man page :is a bit cryptic... SCSI drives have tags, you don't have to lift a finger. Turning on write caching on top of using tags is extremely dangerous anyway, much more dangerous then turning on write caching for an (untagged) IDE drive. ::: net.inet.tcp.keepidle=10000 ::: net.inet.tcp.keepintvl=10000 This will be irrelevant for the benchmark they ran. It is related to keepalive timeouts that terminate dead TCP connections. :What do these do? : ::: vm.pageout_algorithm=1 : :Ditto...? ;-) Don't mess with that, it's magic :-). No, actually, changing it just changes the pageout algorithm to strict LRU, which you almost never want to do. It will not have any effect on the benchmark since they were not stressing memory and it would probably be detrimental if they were stressing memory. ::: ::: as for mount options, "noatime". : :Don't think any Linux distro defaults to noatime mounts, and in Win2K, :you have to hack the Registry to disable Last Access Timestamps. : :-- Juha atime/noatime would have no effect on the benchmark anyway. Or not much of one, anyway. -Matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200107122159.f6CLx3M51540>