* Gaudenz Steinlin (gaudenz@soziologie.ch) wrote: > Am Die, den 20.01.2004 schrieb Stephen Frost um 21:58: > > Alright, first a few bitches: > > - Waiting for DHCP to time out sucked, I know it won't work, > > eth0 isn't even connected to anything > > - Couldn't configure my wireless card (eth1), though it did > > appear to be detected and looked operational > If it was really operational, then you would have been presented with a > list of interfaces. This could be either a bug in netcfg or the > interface is not listed in /proc/net/dev and probably not operational. > Please send us the content of /proc/net/dev at this stage. I'll see about doing that tonight, it did show up as eth1 in ifconfig though. It's a wireless card though, I don't know if that has some impact on this. > > - bash on con2, for whatever reason, stops responding if you hit > > 'alt-1' by accident (like when going for alt-f1 to get back to > > the installation), this is *very* annoying since you have to > > restart the install to get a shell to work in, it'd be nice to > > have a way to forcibly reset the shell or bring up another one > > or something > The shell is busybox ash. This is probably either a kernel bug or a bug > in the busybox shell. Don't know more about that, sorry. It's really annoying. :) I'm working around it though. It's still be nice if there was a way to kill/restart the shell on that console somehow though, or start up another shell on another console. > > - Every other step it appears to 'detect hardware' or whatever, > > what's with that? I don't have a floppy on the machine but > > every time it takes a few seconds to try and load it. > The hardware detection is a bit hackish atm. I don't know of a good way > to detect a floppy drive other than trying to load the module. We should > optimize the code to only try this once. Right, do it once and cache the info, don't redo it every time, that's silly. > > Now, the big problems: > > - discover just hung for a *long* time (>30 minutes), it was > > called with 'modules-detect all', iirc. This was in the > > postinst of discover after the 'base system' had been > > installed and it was installing 'extra components', iirc. > > I didn't see any kernel oops or anything abnormal, and I was > > able to kill discover off (though it just restarted and got > > stuck at the same place). Eventually I just hacked up the > > postinst to act like it didn't find anything (MODULES=""). > > This got me through the rest of the install and to reboot. > Do you know which module it tried to load when it hung? I have never > seen that before, but the most probably explanation seems that it tries > to load a kernel module which hangs. I tend to agree, but I have no idea which module it was trying to load.. Does it log which module it's trying to load something? It really should, if it doesn't... Then you could just flip to con-3 or whatever one is the tail -f and see. > > - On boot, everything looked alright except after saying > > something like 'lp0: detected blah blah' nothing else showed > > up. It looked to have just hung at that point. Maybe it > > was trying to load up some framebuffer or maybe X is supposted > > to start at that point? I dunno, but I couldn't do anything, > > none of the other consoles appeared to be up or anything. I > > recall being warned that there might be a problem with the > > current X in Debian and the graphics card in my laptop (ATI > > Radeon IGP 345M 64MB (shared)), though 4.3 (released *how* > > long ago?) supports it. Don't know if that's a problem here > > or not. > Probably the same discover bug. Yeah, I did find out though that I can ctrl-c and it dropped through and I finished the install and the system appears to be operational, though I can't seem to find drivers for the installed kernel for my wireless card (even though I know it showed up as eth1 in the boot system, and I think I recognized the module in the install, I'll have to check that again). Thanks, Stephen
Attachment:
signature.asc
Description: Digital signature