From mboxrd@z Thu Jan 1 00:00:00 1970 Message-Id: <200306241819.h5OIJB705299@augusta.math.psu.edu> To: 9fans@cse.psu.edu Subject: Re: [9fans] more silly fossil questions In-Reply-To: Your message of "Tue, 24 Jun 2003 19:08:37 BST." From: Dan Cross Date: Tue, 24 Jun 2003 14:19:10 -0400 Topicbox-Message-UUID: da080f4e-eacb-11e9-9e20-41e7f4b1d025 > I reached exactly this stage a couple of months ago and my question > was answered by people going, yes, build your kernel to use > fossil/venti just as it says in the wiki. It may say there that is is > complicated, but that is because it is more complicated than booting > off a CD, not because it is challenging. > > The recipe given in the wiki is pretty easy to follow. > > Once you've got that up and running, everything just behaves like you > want it to. Apart from the fact that changes in your configuration > and pulling updates becomes a bit of a hassle, as you have to > recompile, reinstall your kernel and reboot. I haven't yet gone > through this process myself. But considering all the changes of late > I probably will have to soon. I suppose another solution is to dedicate a (very) small partition at the beginning of the disk for the fossil/venti bootstrap. Put a simple filesystem, such as the flash FS, on it and then the kernel image need only contain just enough logic in /boot/boot to mount it and run a script. It would certainly make alterations easier as they are no longer tied into the not-easily-mutable kernel image. If something *really* gets messed up, one can boot off a CD and ostensibly still have access to the configuration data for recovery. That said, I haven't tried it. I have too much of the wrong hardware to set up a fossil at the moment. With luck, that'll change after I get back from India. - Dan C.