Date: Tue, 04 Mar 1997 13:02:36 -0500 From: Michael Alwan <alwan@rma.edu> To: Nadav Eiron <nadav@barcode.co.il> Cc: questions@freebsd.org Subject: Re: disk usage reporting Message-ID: <3.0.1.32.19970304130236.00690c6c@rma.edu> In-Reply-To: <331BE9E6.24C@barcode.co.il> References: <331BBB00.41C67EA6@rma.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
At 11:22 AM 3/4/97 +0200, you wrote: >Michael Alwan wrote: >> >> To all: >> >> Using df to look at my disk space, I get some confusing reports: >> >> free capacity mounted >> on >> /dev/wd0a -1282 104% / >> /dev/wd0s4f etc etc > >This is a *very* weird setup. Why mounting something on /etc? My apologies for the confusing shorthand. "Etc" just meant more output I didn't think was related to my question. These are my actual file systems: 1-k blocks used avail capacity mounted on /dev/wd0a 31775 29903 -603 102% / /dev/wd0s4f 371087 126120 215281 37% /usr /dev/wd0s4e 29727 14061 13288 51% /var procfs 4 4 0 100% /proc >> Is the wd0a file system mentioned above >> the root file system, and is it filling up because I log on as root and >> all my mail and web caches are there? Why is it /dev/wd0a and not >> /dev/wd0s4a? Perhaps these are the wages of the evil sin of logging on >> as root--perhaps if I had a /usr/home directory I wouldn't be running >> out of space. Any suggestions on how to handle this problem in the >> future (relocate mail and disk caches--a temporary fix?), are my >> suppositions correct, and what about the file system name >> discrepancies. Any light shed on these matters would be appreciated! >> >> Thanks, >> >> Michael >The root file system is reported without the slice number for historical >reasons. Anyhow, FreeBSD will always boot the first BSD slice off the >hard drive, so the slice number for the root filesystem is implicit. >Using the root filesystem for all your needs is not recommended. Apart >from getting it filled up you may also get it corrupted. The standard >setup (with /usr and /var on separate file systems) will ensure that >under normal operation the root filesystem is read-only, lowering the >risk of corruption. > >Running as root also lets you use the space that UFS normaly reserves to >allow faster allocations (that's why you have negative free space on >it). This has a very large impact on write performance. In short - I >usually try to use the root filesystem as little as possible. Most of my >machines have just 20MB of root filesystem, and I never had it filled >up. > >Your df output doesn't show any additional filesystems, but is /usr on a >separate partition? If not, moving to /usr/home won't help... > >Nadav > So, installing Netscape Communicator, xfmail, and all those other goodies while running as root has dumped their caches and so on in the root ( / ) filesystem--correct? Besides getting these caches out of root, what other kinds of cleanup should I do? I can change the cache locations from Netscape, and prefer not to have to reinstall everything if that is realistic. But I get the impression that installing so many things as root may have totally screwed up my file systems in ways I don't yet understand. I assume the correct thing to do is create at least one user and user group like the setup program suggested. I had some problems with this before--specifically, creating a user that had permission to su to root so I could work on configuration files. I tried to create a user in the group wheel and was told wheel was not a valid group assignment. Could you give me a couple pointers to keep me from destroying my system until I finish RTFM? :) Thanks, Michael
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3.0.1.32.19970304130236.00690c6c>