Hi I set up a symlink /var/log -> /boot/log where /boot is a reiserfs system. That was because Reiser4 lost log data when we were chasing the last bug with VS.
Now the partition grew full and it was shown by my computer completely locking up. Out of the blue. At least that's what I assume caused the lock-up and the following. On a reboot I got a trace on the screen from dmesg dying. Unfortunately I can't use netconsole for some reason, but it was there. Sometimes it boots but screams a lot that the logfiles are not found, and for some reason /boot never mounted. That's mysterious because /boot must be mounted for the kernel to be loaded, I don't have a copy in the mount point directory, but dmesg says nothing about anything but Reiser4 being mounted on / I don't frankly have the time now to reproduce this, it is somewhat sporadic and I tried it too many times already. The system is a Debian Sid (Unstable) and the command was dmesg -s 524288 > /var/log/dmesg on a full reiserfs /var/log -> /boot/log partition/directory. Now that I got my system up, it did not mount /boot or /home/mjt, only / and it also lost the host name. (none):/etc# is my prompt. The filesystem is 128480 blocks with just as many used. Hope this helps in any way! PS. Now that I kicked in a reboot, I'm getting: reiser4[run-parts(2707)]: traverse_tree (fs/reiser4/search.c:755)[nikita-1481]: WARNING: Too many iterations: 1048576 Where the iteration count varies. Warnings are usually not lethal, right? But this seems to end never. Time for the sync-umount-reboot sysrq magic.. Thanks! -- mjt
