[systemd-devel] Errorneous detection of degraded array

[Reindl Harald]

Am 30.01.2017 um 07:36 schrieb NeilBrown:
>> By virtue of "Following" attribute. dev-md0.device is Following
>> sys-devices-virtual-block-md0.device so stopping the latter will also
>> stop the former.
>
> Ahh.. I see why I never saw this now.
> Two reasons.
>  1/ My /etc/fstab has UUID=d1711227-c9fa-4883-a904-7cd7a3eb865c rather
>     than /dev/md0
>     systemd doesn't manage to intuit a 'Following' dependency between
>     the UUID and the mount point.
>  2/ I use partitions of md arrays: that UUID is actually /dev/md0p3.
>     systemd doesn't intuit that md0p3.device is Following md0.device.
>
> So you only hit a problem if you have "/dev/md0" or similar in
> /etc/fstab.

the fstab in the virtual machine which ad that issue around 3 times in 
the last year is using UUID - so it's not hit often at all but wouldn't 
have expected it especially hit a VM (reason for that raid is just 
testing RAID10 with writemostly option)

UUID=f0b27a5c-7e3d-45ad-8b7f-617820860379  /mnt/raid10 ext4 
defaults,commit=30,inode_readahead_blks=16,relatime,lazytime,barrier=0  0 1