Date: Fri, 12 Apr 1996 10:53:47 -0400 (EDT) From: surfmwb@nhr.com To: alex@fa.tdktca.com Cc: hackers@Freebsd.org Subject: Re: RE: unable to mount / Message-ID: <199604121453.KAA05330@geneva.visi.net>
next in thread | raw e-mail | index | archive | help
> Mark> I currently have release 2.0 loaded on a system I use for an > Mark> internet email server. All of a sudden today while a remote > Mark> session was logged in trying to telnet out to another it > Mark> reported no available network sessions available..while > Mark> being the only one logged in... > > Mark> Then I went to the console to log in and no matter which of > Mark> my logins including root would allow access. All reported > Mark> invalid username. > > Mark> Unfortunately, I had no other option that I know of so I > Mark> powered down the system thinking I could then fsck the > Mark> system and go about my business..instead the kernel runs > Mark> properly until it reaches mounting the /. It reports the / > Mark> was unmounted improperly and locks up. > > As Joerg pointed out, you can try the fixit floppy. > > Mark> What can be done without a reload, if anything? An what can > Mark> I do to make the reload easier of Release 2.1? I've never > Mark> done a complete load before. > > If you look on the bright side, upgrading from 2.0 to 2.1 is a very > wise move. I've heard some very bad things about 2.0. As for making > the reload easy -- you needn't do anything, it's very easy. I would > recommend that you create separate paritions for different mount > points. For example, I have: > > /dev/sd0a on / (local) > /dev/sd1s1e on /tmp (local) > /dev/sd0s1d on /u1 (local) > /dev/sd1s1f on /u2 (local) > /dev/sd0s1e on /usr (local) > /dev/sd0s1f on /var (local) > /dev/sd0s1g on /var/mail (local) > /dev/sd0s1h on /var/spool (local) > > The benefits are: > > - Spool areas are separated. If I get mail bombed, I can still > print. If /tmp fills up, I can still get mail, etc. > > - Areas such as the root file system rarely get touched, so > if the system goes down, there's less chance of corruption. > > - If a file system does get corrupted, the damage is contained > within a smaller area (for example, I had /u2 go bad last > week, but /u1 survived intact). > > - You can run a filesystem you don't care about (such as /tmp) > in async mode for better performance. (UFS defaults to sync > metadata for safety purposes.) > > Alex > I'm in the process of trying to use the fixit floppy. It boots to the Boot: then it can't find the kernel. What am I doing wrong or what can I do at this point? The kernel is found if I'm trying to boot the system normally but it freezes when mounting the root. Thanks
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199604121453.KAA05330>