From owner-freebsd-chat Mon May 13 13:48:30 1996 Return-Path: owner-chat Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id NAA21110 for chat-outgoing; Mon, 13 May 1996 13:48:30 -0700 (PDT) Received: from brasil.moneng.mei.com (brasil.moneng.mei.com [151.186.109.160]) by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id NAA21101 for ; Mon, 13 May 1996 13:48:26 -0700 (PDT) Received: (from jgreco@localhost) by brasil.moneng.mei.com (8.7.Beta.1/8.7.Beta.1) id PAA05730; Mon, 13 May 1996 15:34:06 -0500 From: Joe Greco Message-Id: <199605132034.PAA05730@brasil.moneng.mei.com> Subject: Re: 3 terabytes on one server? (was Re: more than 32 scsi disks on a single machine ?) To: taob@io.org (Brian Tao) Date: Mon, 13 May 1996 15:34:05 -0500 (CDT) Cc: freebsd-chat@freebsd.org, asami@cs.berkeley.edu In-Reply-To: from "Brian Tao" at May 13, 96 04:24:07 pm X-Mailer: ELM [version 2.4 PL24] Content-Type: text Sender: owner-chat@freebsd.org X-Loop: FreeBSD.org Precedence: bulk > On Mon, 13 May 1996, Joe Greco wrote: > > > > Be thankful our newfs is _fast_... :-) the one in Slowaris is > > slooooooooooooooooooooooooooooooooooooooowwwwwwww... > > Even Sun rates their fsck at "10 minutes per gigabyte"... that > would mean a 3-hour fsck on my news server with Solaris. :-/ OTOH, > they have ODS 3.0 with journalled file systems which pretty much > eliminates boot up delays caused by inconsistent filesystems. They also have PrestoServe, which doesn't incur an additional speed penalty on news filesystems... :-) I'm not overly impressed with the SCSI I/O on a Sun. It "feels" slow. > > There is definitely an advantage in parallelism within your > > filesystems. It is an impressive sight to see news.sol.net first fsck > > /, /usr, and then /var, and then watch a dozen drives suddenly chime > > in with a massively parallel fsck :-) > > Waitaminute.... does it do this by default? It takes 10 minutes > to fsck the news server filesystems, and I'm pretty sure it only does > one drive at a time. It would be great if multiple fsck's could run > in parallel. Maybe my /etc/fstab isn't setup correctly? > > /dev/sd0a / ufs rw 1 1 > /dev/sd0s1e /usr ufs rw 1 1 > /dev/sd0s1f /var ufs rw 1 1 > /dev/sd1s1e /usr/local ufs rw 1 1 > /dev/sd2s1e /var/spool/news ufs rw 1 1 > /dev/sd3s1e /var/spool/news/alt ufs rw 1 1 > /dev/sd4s1e /var/spool/news/alt/binaries ufs rw 1 1 > /dev/sd5s1e /var/spool/news/over.db ufs rw 1 1 > /dev/sd6s1e /var/spool/news/comp ufs rw 1 1 > /dev/sd6s1f /var/spool/news/soc ufs rw 1 1 > /dev/sd7s1e /var/spool/news/misc ufs rw 1 1 > /dev/sd7s1f /var/spool/news/rec ufs rw 1 1 > /dev/sd8s1e /var/spool/news/out.going ufs rw 1 1 > /dev/sd8s1f /var/spool/uucp ufs rw 1 1 > > /dev/sd0s1b none swap sw 0 0 > /dev/sd2s1b none swap sw 0 0 > > proc /proc procfs rw 0 0 Uh, buddy, RTFM!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! :-) :-) :-) hummin# cat /etc/fstab /dev/sd0a / ufs rw 1 1 /dev/sd0s1e /usr ufs rw 1 1 /dev/sd0s1b none swap sw 0 0 /dev/sd1s1e /var ufs rw 1 1 /dev/sd1s1f /var/spool ufs rw 1 1 /dev/sd2s1e /usr/local ufs rw 1 2 /dev/sd12s1e /nfs ufs rw 1 2 /dev/sd13s1e /nov ufs rw 1 2 /dev/sd10s1e /nov/.0 ufs rw 1 2 /dev/sd22s1e /news ufs rw 1 2 /dev/sd14s1e /news/.0 ufs rw 1 2 /dev/sd15s1e /news/.1 ufs rw 1 2 /dev/sd11s1e /news/.2 ufs rw 1 2 /dev/sd21s1e /news/.3 ufs rw 1 2 /dev/sd20s1e /news/.4 ufs rw 1 2 proc /proc procfs rw 0 0 I will leave the difference to the reader to find. Note this doesn't include my ccd partition which I set up and fsck separately because I was too lazy to go and integrate it. It's almost fun to watch it fsck, except the fact you have to fsck implies you crashed, which _isn't_ fun. ... Joe ------------------------------------------------------------------------------- Joe Greco - Systems Administrator jgreco@ns.sol.net Solaria Public Access UNIX - Milwaukee, WI 414/546-7968