I had a raid card failure on a box in Texas (thanks HP). They replaced
the card, but now I need to fsck to ensure the fs's (which seem ok) are
not messed up. Should be fine because they are journalled, but best to be
sure.
On the root fs obviously I'll have to force after reboot. But I want to
make sure it doesn't stop to ask me anything, as no one is there, and for
someone to look at it would take hours.
Anyone know the best way to achieve that? Does fsck on reboot (using the
magic file method) run with -y, or what?
I found /etc/e2fsck.conf and man e2fsck.conf, but the only hints I got
from there regarding non-prompting is possibly the "problems stanza", and
possibly listing all/most problems manually and setting preen_ok=1 on
each? Seems like a lot of work, especially given the warning in the docs
in that section. And the docs don't list all the codes, so it looks like
one has to dig in the source.
I suppose I could be happy with an option that says auto-yes to all the
minor/moderate problems but still can blowup waiting for console input for
disastrous problems.
What I don't want is for the thing to sit and prompt just because some
dir is missing an inode or an extent isn't quite the right length.
Thanks!
P.S. here's what I've come up with so for for a conf file with options
that sound useful to me:
[options]
broken_system_clock = 1
inode_count_fullmap = 1
log_dir = /boot/tec-fsck-logs
log_dir_fallback = /root/tec-fsck-logs
log_dir_wait = 1
log_filename = e2fsck-%N.%h.INFO.%D-%T
max_count_problems = 65536
readahead_mem_pct = 5
report_verbose = 1