Date: Tue, 03 Feb 1998 09:23:22 -0800 From: Bill Trost <trost@cloud.rain.com> To: John Goerzen <jgoerzen@alexanderwohl.complete.org> Cc: mobile@FreeBSD.ORG Subject: Re: Hard drive not spinning down Message-ID: <19980203172424.6694.qmail@jli.com> In-Reply-To: Your message of Tue, 03 Feb 1998 10:05:50 CST. <Pine.BSF.3.96.980203100005.630A-100000@alexanderwohl> References: <Pine.BSF.3.96.980203100005.630A-100000@alexanderwohl>
next in thread | previous in thread | raw e-mail | index | archive | help
John Goerzen writes:
I have been noticing that the hard drive doesn't spin down like it should.
I have already taken the steps of setting async and noatime to all the
filesystems, so mere reads will not cause the hard drive to spin up if the
data is already in its cache.
I doubt that async is gaining you anything -- is I understand it, all it does
is loosen the constraints on the order of disk write operations. It is more
likely to cause you grief the next time you have a crash (say, due to low
battery? NAAAH). I like the noatime idea, though.
I have turned off sendmail's automatic queue runs and the only other things
going were X, fvwm2, and some xterms -- all of which were running telnets
that were logged in to a remote machine.
This isn't a big problem but it is a strange one.
Well, it was something I wanted to discuss anyhow, so thanks for bringing it
up. I was having similar problems, so I went searching the disks for possible
offenders. Things I discovered:
1. History files tend to get written every time I hit enter. Not good. I
squashed some of them, but don't know what to do about the rest.
2. Cron logs (in /var/cron/log) everything it tries to run, and mgetty's
faxrunq was going off fairly frequently (which also writes a file when it
runs). Too bad cron(8) doesn't indicate what priorities are used for what
sorts of messages (I mean, the man pages *I* write tell you absolutely
everything you want to know, they're perfect, uh-huh...).
3. Two modtimes in /dev are updated every time you type a line (character?) at
an xterm. I assume these have to get written out at some point, but FFS might
be willing to let them sit and stew. Maybe laptops should build /dev in an mfs
partition????
If anyone else has ideas on how to keep FreeBSD from touching the disk, I'd
love to hear them. On the other hand, Lose'95 is worse than FreeBSD for me at
this point, so I'm not *too* worried.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980203172424.6694.qmail>