On Fri, Apr 22, 2016 at 1:23 PM, Samuel Sieb <samuel(a)sieb.net> wrote:
On 04/22/2016 12:19 PM, Chris Murphy wrote:
>
> On Fri, Apr 22, 2016 at 12:54 PM, Samuel Sieb <samuel(a)sieb.net> wrote:
>>
>> e2fsck -fn /dev/sda3
>>
>> -f to force it, -n to do it-read-only
>
>
> In the OP's example where an md2 device was affected, it needs fsck
> run on /dev/md2 (assembled and running but not mounted). Since md2 is
> his root fs, that means a.) boot from alternate media or b.) use boot
> parameter rd.break=pre-mount and do it at the dracut shell, but this
> might require manually running the array, I'm not sure whether it will
> be running at this point but it probably won't be mounted (check
> first).
>
Sorry, I was copying the previous poster and forgot the device was specified
earlier.
The command to run is "e2fsck -fn /dev/md2". That can be run while it's
mounted to give an idea of what's wrong with the filesystem.
I would not exclusively trust user space messages. Check dmesg before
and after for the appearance of new messages that may not even be
directly ext4 related. This could be a case of a bad sector
instigating the problem.
--
Chris Murphy