Date: Fri, 14 Jun 2002 13:39:21 -0400 From: Bill Moran <wmoran@potentialtech.com> To: chris <lists@powernet.net> Cc: Steven Goodwin <steve@cit.gu.edu.au>, questions@FreeBSD.ORG Subject: Re: var optimization Message-ID: <3D0A2A49.6010407@potentialtech.com> References: <Pine.GSO.3.96.1020615011850.23729A-100000@kurango.cit.gu.edu.au> <000f01c213c6$844baec0$a701a8c0@reno.powernet.net>
next in thread | previous in thread | raw e-mail | index | archive | help
chris wrote: > This is usually a sign the disk is getting full. It is an automatic thing > that happens when the disk/partition goes over a set amount(90%?) > Check that first.... No, you've got it backwards ... see below. > ----- Original Message ----- > From: "Steven Goodwin" <steve@cit.gu.edu.au> > To: <questions@FreeBSD.ORG> > Sent: Friday, June 14, 2002 8:38 AM > Subject: var optimization > >>Hello FreeBSD users. I received this message immediately after (or as >>part of) my gateway's startup. >> >>login: Jun 14 07:09:41 kaneda /kernel: /var: optimization changed from >>SPACE to TIME Changing from SPACE to TIME indicates that the filesystem is self-tuning itself for better performance. Probably, what happened is that the disk got into a situation where space was running short or it was getting very fragmented. The default optimization is TIME (fastest writes) but when the disk gets messy, the system will change this optimization to SPACE (slower writes, but more organized data for faster reads). Apparently, that condition ended (you deleted some files or over time the SPACE optimization cleaned up the fragmentation problem) so the system switched the optimization back to TIME (which makes for fast reads and writes, as long as there's enough free space to support it) fsck will print out the degree of fragmentation on a filesystem, so you may want to check your logs to see what fsck has been saying at bootup time. If you check your system logs, I'm sure you'll see that at some point in the past, the system switched from TIME to SPACE, and it's now switching back. >>This may or may not be the place to ask this question, but could anyone >>satisfy my curiosity and explain (or point me to some documentation that >>explains) which process or part of the kernel makes this decision and what >>are the reasons/criteria for the change. The tunefs man page has a bit in it about this, as well as the newfs man page. As to which part of the kernel, it's either the ufs or the ffs drivers. I don't know the exact methodology/reasons for the change, the above is a high-level look. There is a doc in /usr/share/doc called diskperf that will probably have some more details on how/why this occurs. -- Bill Moran Potential Technology http://www.potentialtech.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3D0A2A49.6010407>