On Sunday, 17 January 1999 at 22:17:21 -0800, Jake wrote: >> Oops. This was a thoroughly successful commit, wasn't it? At least >> it didn't break `make world'. > > :) > >> >> I'll think out another way and test more thoroughly. How does this >> look? >> >> if [ -f /etc/vinumslices ]; then >> read /etc/vinumslices >> if [ -r /modules/vinum.ko ]; then # jkh paranoia >> kldload vinum >> vinum read $vinum_slices >> else >> echo "Can't find /modules/vinum.ko" >> fi >> fi >> > > Why not just suck in rc.conf before starting vinum?
Good idea. Does anybody have a reason for reading in rc.conf so late? > I've modified my /etc/rc to do this, and it works fine. > I submitted a PR with a patch. I'll look at it. > also, I tried 'pseudo-device vinum' and it does work, > but I think I ran into trouble somewhere. > Oh yeah, it worked once, but each subsequent > reboot seemed to replicate my subdisks, > until I had about 16, only two of which were real. > Finally I found my config file, recreated and switched > back to the kld. I don't think this is the fault of the pseudo-device. This is more likely to be the bug I mentioned in my mail to Matsushita-san. > FWIW I think its good to load the module from /etc/rc > rather than loader.rc, keeps all the vinum stuff in one > place. Maybe we could even have the module loaded > on demand; ie if you try to run vinum before the module > is loaded, load it automatically. Once I have this update out of my hair, I'll think about this. Greg -- See complete headers for address, home page and phone numbers finger g...@lemis.com for PGP public key To Unsubscribe: send mail to majord...@freebsd.org with "unsubscribe freebsd-current" in the body of the message
