fsck during boot no longer shows drives being checked

[doctordruidphd]

All right, this in nit-picky in the extreme, but:

Before 13.10, during boot (I have plymouth disabled) the boot text used to show which file systems were being checked during bootup. Someone has changed fsck to run in a non-verbose mode, so there is no longer any feedback as to what it is doing. I want to change that back, so I can see what drives are being checked. I cannot find where in the initramfs this is being run from, as there is no entry for fsck anywhere in initramfs-tools /hooks or /scripts that mentions it (other than for btrfs). Does anyone know how this gets run? I have searched through just about everything relevant to it in /etc/ and /usr/share/initramfs-tools.

[kubicle]

The thing that runs fsck jobs at boot is the mountall upstart job /etc/init/mountall.conf (that calls /sbin/mountall)

The change that disables fsck messages is likely this one:

mountall (2.50) unstable; urgency=low

* Add support for pstore. Closes: #722212.
* Allow multiple fstab entries with the same mountpoint: we still mask
any entries that came from a different source (i.e., /etc/fstab will
mask /lib/init/fstab), but if there are multiple entries in /etc/fstab,
treat them all as separate mounts instead of letting the last one win.
LP: #503003.
* Don't set 'console output' anymore; now that we have upstart logging by
default, logs will be captured and (assuming /var/log becomes writable
at some point) mountall is debuggable without having to spew text to the
console.
* Support reading --verbose/--debug options from /proc/cmdline, so that
we can give users a better way to debug mountall without having to
edit the upstart job.
* Since everything is in a single event loop, a request from plymouth to
skip a mount may come in after the mount has already happened. So don't
assert on a bogus request, just treat it as a no-op and continue.
LP: #731800.

So one should be able to re-enable boot fsck messages by appending "--verbose" to the linux boot options. You can test this by doing it manually in grub...or you can add it to /etc/default/grub and run 'sudo update-grub' to add it permanently.

Another option would be to edit the upstart job and add "--verbose" to the "exec mountall" -call followed by an initramfs update.

[doctordruidphd]

Thanks. I see that in mountall.conf. Will experiment and report results back.

[doctordruidphd]

Sorry, but this didn't fix it. I tried adding the --verbose option to the linux line in grub.cfg; that certainly produced a ton of output, but nothing about which drive is being fsck'd at boot. Just the message that "drives are being checked" (words to that effect) and "press C to cancel" -- no indication of which drive is being checked, which up until 13.10 it always showed. I also tried adding the -v option to the mountall command in /etc/init/mountall.conf and doing an update-initramfs -u, but still no indication of which drive is being checked. It matters because if it's checking the 1tb backup drive, there's no need to sit there and stare at the screen (it also matters because, why shouldn't I know that?). So not sure where the problem is at this point.

Edit: wonder if the -v should go in /usr/share/initramfs-tools/event-driven/upstart-jobs/mountall.conf ? Will give that a try.

Re-edit: hmm. looks like /etc/init/mouontall.conf got changed back by something (other than me)?? Will have to try that again.

[kubicle]

Having --verbose in the boot options or calling mountall explicitly with --verbose (or -v) options should lead to same results, so I don't think that is your "problem". I haven't checked the source code but the reason might be that /sbin/mountall just doesn't output the information in the console (even with the --verbose option). I agree that is idiotic (if that's the case), but wouldn't surprise me much...ubuntu goes to great lengths to hide useful boot messages from users.

Have you checked the mountall logs in /var/log/upstart/...is there a difference in the logs for --verbose and non-verbose booting.

You can also try the --debug option, although that might include only a lot of debugging messages.

[doctordruidphd]

All of the relevant information is there in mountall.log:

stable: clean, 1053797/6406144 files, 11770885/25600000 blocks
library: clean, 1253955/25346048 files, 78990907/101360384 blocks
kubuntu: clean, 1087896/4161536 files, 11314595/16640000 blocks
aptosid: clean, 931786/6406144 files, 8801416/25600000 blocks
testing: clean, 1054047/6406144 files, 11755745/25600000 blocks
data: clean, 3298281/39804928 files, 100253258/159197696 blocks
fsck / [323] exited normally

So it's being hidden somewhere. Idiotic as you say. Experiments will continue.

[doctordruidphd]

Well, I have tried just about everything I can think of, and still can't get a verbose output from mountall.
I have added the --verbose option manualiy into /etc/init/mountall.conf and /usr/share/initramfs-tools/event-driven/upstart-jobs/mountall.conf (followed by update-initramfs -u), and I have tried adding the --verbose option to the linux boot line in grub2. The first 2 experiments produced nothing; adding verbose to the boot line certainly produces a lot of output, but, it does not produce any indication of what is being fsck'd. Not sure why it's so important that users not know what drives are being checked??

[kubicle]

Well, I have tried just about everything I can think of, and still can't get a verbose output from mountall.

Like I mentioned, it's probably the /sbin/mountall executable "at fault" here (I haven't checked the source so don't quote me on that).

So it might not be possible without editing the source and rebuilding.