Even using vanilla 2.6.0-test4 sources, the system stops when trying to remount the read-only reiserfs root partition to read-write. I tried tracking down the bug, and it appears it is either with /etc/init.d/checkroot or possibily /etc/init.d/checkfs by blanking both files and adding the following to checkroot (yeah, I know it's very ugly and bad, but the system boots now): mount -t reiserfs /dev/hda3 / mount -t reiserfs /dev/hda3 / -o remount,rw This bug has been confounding me since 260-beta1 was in portage.
i can confirm that non root partitions can be remounted rw ...
Can you please make sure you are using the latest baselayout? If the problem persists, can you remove any instances of "&>/dev/null" in the init. scripts and see if that causes anything of interest to be displayed.
I had the latest non-masked ebuild of baselayout,sys-apps/baselayout-1.8.6.8-r1, but ran a "ACCEPT_KEYWORDS="~x86" emerge baselayout and got updated to baselayout-1.8.6.10, which seems to have the problem fixed. I am guessing it has something to do with the way checkroot stores and retrives the value for the root filesystem. Btw, before I upgraded the baselayout, I got the output of checkroot, here it is if anyone is curious: mount: / not mounted already, or bad option
please don't kill me ;) The actual problem was in my fstab - apparently, 2.6 does not like when you set your root parition options to "tail", it works fine with "notail" though. When I update my baselayout, my fstab got overwritten, but then when I started changing things back, I found I was having the same problem again.