Bug#788050: systemd-fsck : Check disks at each reboot

[Allen Webb]

I edited the /etc/default/grub file, removed splash from 
"GRUB_CMDLINE_LINUX_DEFAULT" and ran update-grub2. To force the fsck I used:
sudo tune2fs -C 30 <device>
and had to reboot twice. I saw the fsck progress, but it still stopped 
at the same time and gave an operation error / sigpipe.

On 09/23/2015 12:33 PM, Michael Biebl wrote:
> Am 23.09.2015 um 18:08 schrieb Allen Webb:
>> On Fri, 14 Aug 2015 10:13:25 +0200 Axel Ludszuweit
>> <axel.ludszuweit at htp-tel.de> wrote:
>>   > Dear maintainer,
>>   >
>>   > I also think there is is a timeout problem.
>>   > I Use lvm2 with md software raid based on ext3.
>>   > Systemd file system checks at boot-up on small partitions finish
>>   > successfull on big partitions end with error 13.
>>   > Manual executed file system check on these big partitions end
>>   > successfull and on the next boot-ups no errors appear until new fsck is
>>   > triggered by reaching boot count limit.
>>   >
>>   >
>>   >
>>   >
>>
>> I am running into this same issue downstream testing ubuntu 15.10. My
>> small partitions check fine, but one that takes a lot longer to check is
>> always fails the check during startup. I thought the problem might be
>> related to either dm-raid or a timeout, but if other people are having
>> the problem it is probably a timeout or race condition between other
>> services. When I run a manual fsck.ext4, I have no issues at all so it
>> seems to be an integration problem with systemd.
> We suspect that this is related to systemd-fsckd, a Debian/Ubuntu
> addition to provide fsck progress feedback for plymouth.
>
> If you boot without "splash" on the kernel command line, does it make a
> difference?
>
>