I have problems with initng again.

Sometimes, every 2 boot or so, initng finds that checkfs took too long to load and decides to kill the service, resulting in a hanged computer.

How do I override that timeout for checkfs so my root partition loads correctly?

thanks

I've had a similar problem. On my box, the problem seems to be with non-root partitions . . . I've looked at the mountroot script, to see what it does differently from mountfs, and the most relevant thing seems to be that, for non-root partitions, system/checkfs and system/mountfs are different processes, whereas the check script for the root partition is part of system/mountroot: system/mountroot/check. Perhaps, by including the check-the-filesystem script in the mount file, before the actual mount script, it forces initng to wait for any scans of the root partition to complete.

But that's only an uneducated guess. I'd love to find out the answer to this, since it's one of the reasons I'm inclined not to use initng.

It has stabilized recently. It did act up twice when I added another partition, hda8 (for qemu), but haven't heard problems of it since (nor have I rebooted it since the last successful boot, huh).

I think this worked for me: put " critical; " before " never_kill; " in system/checkfs. The fsck root and mount root parts of system/mountroot are marked critical, as is mountfs, but checkfs isn't. At least, for me it wasn't.

um, I'll try that

I was thinking, perhaps it's udev or something. In my terminal, all the output is scrambled at that point, and FAIL may either belong to checkfs, or udev, which seams to have spewed a few outputs before checkfs executed...

sure enough, udev doesn't have never_kill, only critical...

or it can be a combination of the two. When two or more things don't work properly, we have to find what does work properly

the problem is checkfs.

now that it is critical, it gives me a su_login, since it became critical.
It fails at line 54, which is

then
echo "Checking file systems"
logsave -s ${FSCK_LOGFILE} fsck ${spinner} -R -A ${fix} ${force} ${FSCKTYPES_OPT}
FSCKCODE=${?}
if [ ${FSCKCODE} -gt 1 ]
then
echo 1 "code ${FSCKCODE}"
handle_failed_fsck
fi
else

somehow, in there, it fails
However, I don't get the warning messages from checkfs.i, instead, I get the error of initng, smashed with the prompt of the su_login.

I'll google this. If you find the solution, I'm all ears.

HALF-FIXED:
I added those reps:
http://triggerit.tr.funpic.de/blog/?p=42

and installed the latest ifiles. it works now
three boots without problems! whoohoo!

but, the latest initng requires more recent glibc6, which I don't have.
I might just use the source instead eventually.

nevermind