Hi,
I maintain several linux boxes, some of them doesn't have a kbd nor
a monitor connected to them, some of them are difficult to reach.
In order to maintain the with a minimum of manual interaction it
would be nice if they would check *and* repair their filesystems
at boot time - always.
To achieve this one needs to modify /etc/init.d/checkfs.sh alightly.
Instead of
fsck -R -A $fix $force
one should call
fsck -R -A -f -y -p
Since /etc/default/rcS is sourced before this could be controlled
by a variable like "FSCKFORCEREPAIR=y".
What do others think?
Now something different. Tonight I had to reboot a machine where the
filesystem was crashed and several processes including sync were in
status 'D' (non-interruptible). Thus a "reboot" didn't work and
so I tried to omit init and force a reboot with "shutdown -r -n".
Unfortunately this didn't show the expected results. All processes
seemed to be killed (at least my two ssh's were killed, together
with inetd and sshd), so I wasn't able to log in anymore. But the
kernel didn't reboot. This is not funny since I had to ride to their
location and press the reset button.
Is there an easy way to do this remote?
Regards,
Joey
--
There are lies, statistics and benchmarks.
Attachment:
pgpbJomeiziFz.pgp
Description: PGP signature