Re: Skipping fsck during boot with systemd?
On 12/11/2014 03:37 PM, Charlie wrote:
Some of the messages in this thread are an exchange between myself and
Andrei Popescu. Andrei came up with a perfectly easy solution which I
have tested and found to work.
On Thu, 11 Dec 2014 12:23:10 +0200 Andrei POPESCU sent:
The root of my sid install was created before that, so I was still
getting the periodic check for it. The other ext4 filesystems were
newer, so weren't checked (and I didn't even notice it).
I've just disabled the automatic check on the root partition as well,
but I'm considering how to implement a forced fsck every now and
then, including an xfs partition, which wouldn't be checked at boot
I have to admit that I noticed it, but was made an ass off, because I
assumed it was happening in the background. Not realising it always
started before the boot because it couldn't fsck while the machine was
I thought that might have been what allowed systemd to boot faster?
If you find a way "to implement a forced fsck every now and then could
you please post it here. As I would be very pleased to be able to do
that now realising it is no longer happening at all.
Being and "ordinary" user I have no idea where to look or even start.
It involves an edit (as root) of /etc/rc.local to add a single line.
That line invokes the tune2fs utility to set a parameter called maximum
mount count to 0 during the boot process. That keeps fsck from being
For a system which boots from /boot in /dev/sda1 I used:
tune2fs -c 0 /dev/sda1
as the line added to rc.local.
When I want to manually run fsck on that partition I use:
# tune2fs -c 1 /dev/sda1
That sets maximum mount count to 1, which means that fsck will be run on
/dev/sda1 the next time the system is booted. This strategy means that,
following the file system check, rc.local will again use tune2fs to set
the maximum mount count back to 0 so that fsck won't run on the
And Bob's your uncle!
another ordinary user -- JP