Am Tue, 27 Mar 2007 12:53:41 +0200 schrieb Sebastian Vahl ml@deadbabylon.de:
Am Tue, 27 Mar 2007 11:21:13 +0200 schrieb Sebastian Vahl ml@deadbabylon.de:
Ok. There is no package missing. But I have no idea what is getting wrong. If I chroot into the new installed partition (no matter if from normal installation or from livecd) and reinstall the kernel, this installation boots then. If I compare the extracted initrds before and after reinstalling the kernel the only difference is, that the last one doesn't contain dm-mirror.ko, dm-mod.ko,dm-snapshot.ko, dm-zero.ko (I'm not using lvm on my host). And in the init itself sdb1 is replaced with LABEL=/1 in mkrootdev.
Is there a way to tell liveinst/anaconda to be very verbose? --loglevel=debug inserted in /usr/sbin/liveinst does not produce interesting logs.
Here is a "screenshot" from the error. The next point there normally should be to start "Red Hat nash": http://img261.imageshack.us/my.php?image=p3210195gi7.jpg
I've also tried a two days old version without the groups in kickstart file. This one does not show the error and that confuses me. Maybe the error is produced by a package that do not exist in old version?
It seems to be a selinux thing. When installing with selinux enabled there are no entries in auditd.log. But booting from the new installation is not possible. When booting the livecd with enforcing=0 there are still no entries in audit.log but the new installation could boot.
With the first installation I've tried to recreate the initrd inside a chroot into this partition. when using /sbin/newkernel-package, setroubleshoot gives me the attached error. After doing an "restorecon -v /lib/ld-linux.so.2" (inside chroot) I don't get the error and the new created initrd is working.
I will check my hostsystem. The error must be there, I think.
Sebastian