After restarting system after a power outage, or hardlock a root filysytem of type reiserfs will not boot properly. This is becuase the way reiserfsck is called currently does not correct errors on filesystem. The file /etc/init.d/checkroot calls: fsck -a / which calls fsck.reiserfs -a /dev/hda{x} From the reiserfsck man page: -a, -p These options are usually passed by fsck -A during the automatic checking of /etc/fstab partitions. For compatibility, these options simply cause reis- erfsck to print information about the specified file system. No checks are performed. When it is set - reiserfsck assumes that it is called by fsck -A, provides some information about the specified filesystem and exits. -V This option prints the reiserfsprogs version and exit. -r, -p, -y These options are ignored. -V, -f prints version and exits In order to check a reiserfs system the calls need to be something like: echo "Yes" |fsck -C -V -a / -- --check if [ "$?" -eq 0 ]; then continue boot elif [ "$?" -eq 1 ]; then echo "Yes" |fsck -C -V -a / -- --fix-fixable elif [ "$?" -eq 1 ]; then fsck -C -V -a / -- --rebuild-tree fi
that last elif should check [ "$?" -eq 2 ];
uhh, reiserfs doesn't need checked after a hard reboot, that is the point of a journalling filesystem...
i thought they still needed checked if not cleanly unmounted. jounalling just makes the consistency checks faster. the problem i have been seeing after an unclean reboot is that either a bunch of garbage or a bunch of bogus module errors scroll up the screen. i have to <ctrl><alt><del> while that is going on to cleanly reboot. the next time it comes up it is fine because the filesystems where cleanly unmounted. -a <man page quote> simply causes reiserfsck to print information about the specified file system. no checks are performed. </man page quote> hence no journal logs are replayed to ensure filesystem consistency (but i could be wrong about this, and the issue is something else) i had this issue with my 1.2 install as well....
WRONG. When you mount a journaling filsystem part of the mount process is to replay any journal entries who's data modifications haven't already been made on the physical disk.
Haha I was just about to post I was wrong. I was able to see the log replay just before init is called. Maybe there are still issues with reiserfs? The error behavior seems to point to that. This only happens when X locks....
*checks namesys page* I crash my box a lot with overclocking and all sorts of goodies... and the fact that I'm constantly hacking kernels together and I have all my filesystems on reiser and have never experienced this.
I just installed a brand-new Gentoo from the bootstrap CDs. My root filesystem is reiserfs. Gentoo's initscripts say "Checking root filesystem..." and then bomb out on fsck.reiserfs: unrecognized option '-a'. I tried adding 'nocheck' as an option to my fstab. This did nothing. So, is there an actual solution to this problem? How do I make Gentoo not try to check my root filesystem? Is the accepted method to edit the initscripts manually?