/proc mounted readonly.cant rmve /run/openrc[solved]hddcrash

[Mgiese]

hi there,

some month ago i set up an office pc for my uncle. it was working flawlessly until some weeks ago.

when he tried to login gdm there was an error that he cannot write and so on.....

by no means could he have filled the harddisk. but to be on the safeside i just deleted some hundred megs, but it still did not change.

then i did : df

and df tells me : rootfs as well as /dev/root are 100% full ??????!! (even after deleting 100 megs or more)

when i restart the pc i can see one error message when mounting /proc.
saying : cannot remove /run/oprnrc.

can someone tell me what might cause this strange and unexpected behaviour ??


thanks a lot !!

edit : when restarting the system i get a message : remounting / read-only.

why and what ???? nobody except me had root(so nothing/nobody should have access to important config files and so on), so how can it be possible that the system got screwed up ??

[NeddySeagoon]

Mgiese,

root can fill up with log files is /var/log/* When root fills up, its very painful as logs and lock files cannot be written.
The normal startup sequence mounts root readonly, runs rootfsck and if its OK, root is remounted rw.
It appears that rootfsck is failing.

Save everything you can from root, then try a fsck on it using a liveCD.
fsck can make things worse, hence the backup before you start.

[Mgiese]

i am positiv that the harddisk is not full. the pc was rarely used and it had no services running that could create tons of log files to fill up a 200gb hdd.

i deleted lots of stuff. /usr/portage/distfiles 1,5gb as example. but as i wrote before, the condition of 100% filled /dev/root did not change, not even after restart.

i could not check if fsck detects any errors on the FS, because i had no livelinux with me yesterday at the location the machine was working. i will try your suggestions within the next 3 days.

thanks a lot


edit : i completely forgot to mention that at first start the FS was being checked (every 30 days) and fsck was standing still for some time at 24,4%. that makes it very likely that the harddisk may be damaged , am i right ??

[NeddySeagoon]

Mgiese,

Post tho output of

Code: Select all

df

and

Code: Select all

df -i

please.
I suspect that root is corrupt, rather than full.

[Mgiese]

hi,

now i used a live linux with fsck -c (scan for bad blocks) and at the end i got the message "Filesystem WAS MODIFIED". the system is booting up again.

so what could have caused the corruption ?? a simple hard switch of (power off) ? or is it more likely that the harddisk crashed ??

is there a way to show the "bad blocks" ??

Code: Select all

~ $ df
Dateisystem    1K-Blöcke Benutzt Verfügbar Verw% Eingehängt auf
rootfs         112950756 6325784 100887360    6% /
/dev/root      112950756 6325784 100887360    6% /
tmpfs             445696     380    445316    1% /run
udev               10240       0     10240    0% /dev
shm               445696      72    445624    1% /dev/shm
cgroup_root        10240       0     10240    0% /sys/fs/cgroup
localhost ~ $ df -i
Dateisystem     Inodes IBenutzt   IFrei IUse% Eingehängt auf
rootfs         7176192   352221 6823971    5% /
/dev/root      7176192   352221 6823971    5% /
tmpfs           111424      403  111021    1% /run
udev            111424      428  110996    1% /dev
shm             111424        3  111421    1% /dev/shm
cgroup_root     111424        6  111418    1% /sys/fs/cgroup

[NeddySeagoon]

Mgiese,

Yes, a hard power off can corrupt any filesystem mounted r/w.

The hard drive remaps blocks that are getting difficult to read and moves the data, so unless something fails suddenly, (it happens) the drive hides bad blocks from the operating system. The SMART data tells how many sectors have been remapped already and how many more the drive would like to remap.

New drives already have some remapped sectors, which is how they appear to have no defective blocks.

emerge smartmontools and try smartctl -x /dev/sd...