On Wed, 2002-10-09 at 09:28, Guillaume Cottenceau wrote: > Ryan S Oltman <[EMAIL PROTECTED]> writes: > > [...] > > > > > tried just about every combination possible and still no luck. Is there > > > > some step I need to be sure and do in order to guarantee the system is > > > > patched? > > > > > > No, it should work. I've tested it.. > > > > Maybe because I have to pass more kernel options (ide2=... ide3=...) it > > doesn't work? Also, I had to download the patch.pl using a windows PC > > it could be a <CR><LF> problem? It does go and act like it loads > > I don't think so.. maybe the error could come from the fact that > you have more than one RAID array (Pixel, do you think your patch > should work with 2+ RAID array's?). > > > something off the floppy, but I get the same result when I reboot. The > > logs say "PATCHED" during this stage. > > So at least the patch seems applied. It should work :-(.
I would get an error like "no patch-oem-.pl" when I tried to patch it off CD1 with the floppy, but I got the "PATCHED" statement when I booted off CD2 then installed CD1. > > > Anyway, I removed the "/" from raided partition and broke it up into /, > > /opt, and /var all 2.7 Gb in size. I then rebooted off the rescue CD > > mounted the partitions and fixed the /etc/raidtab. It was as follows: > > > > <<<original>>> > > /dev/md3 > > device /dev/hdg6 > > Hum this looks like the contents of an /etc/raidtab *without* the > patch??.... > > [...] > > > The machine now boots! (small victory!). Now I have found that neither > > urpmi nor rpmdrake seem to work. The processes seems to hang w/o any > > CPU activity. I think I found a thread to fix this issue. > > The rpm db is probably locked. Double check you have no rpm > process, then do "rm -f /var/lib/rpm/__*" and "rpm --rebuilddb". > This is what I did and it fixed it. > > Since I installed with out the PCMCIA controller installed, I assume > > I'll have to install: > > pcmcia-cs-3.2.0-3mdk.rpm > > pcmcia-cs-x11-3.2.0-3mdk.rpm > > In order to get the card to work. > > Humm there is also the detection step.. I don't know if harddrake > can do that after the install. No I had to install the packages. The /etc/sysconfig/pcmcia file needed to be modified to: PCMCIA=yes PCIC=i82365 Then, /etc/rc.d/init.d/pcmcia start It detected the card and worked with wvlan_cs driver. I then configured the network, but was not allowed to setup internet connection sharing because it only said I had the eth0 (linksys card). Anyway, when I rebooted the fun started. On the reboot the during the pcmcia startup the messages said something like: "loading yenta_socket instead of i82365". I have never seen the yenta_socket driver on my machine. The controller beeped twice, then it did a stack dump on only one of my processors (I didn't know this was possible). About 5 seconds later the entire machine hung. Long story short, I downloaded the pcmcia_cs-3.1.2 from sourceforge, recompiled the kernel enterprise configuration without pcmcia support and installed the sourceforge package. The machine is functioning now with the i82365 module and is running great. As someone stated in a different thread, I am confirming you must do a make mrproper in order to compile the mandrake kernel source. > > > -- > Guillaume Cottenceau - http://people.mandrakesoft.com/~gc/
