Re: mdadm and fsck
Hi,
Joey L wrote:
I have a broken raid1 drive set - i was able to fix 1 but the other raid
still has issues.
You fixed the file system with fsck. One drive [of md0] is considered
fine, the other is becoming fine during the rebuild.
i have gone through the tutorials online to bring it back online but
still unable to bring it back.
Do i have to run fsck on the individual member drives ???
No.
my /proc/mdstat looks like:
Personalities : [raid1]
md1 : active raid1 sdc1[0] sdd1[1](S)
976758841 blocks super 1.2 [2/1] [U_]
Not sure, but it _may_ be that md1 will rebuild itself using the
"*S*pare" when md0 is fully rebuilt.
md0 : active raid1 sda1[2] sdb1[1]
972654456 blocks super 1.2 [2/1] [_U]
[==========>..........] recovery = 50.7% (493690944/972654456)
finish=416.5min speed=19162K/sec
..
my questions is :
1. how can i bring it back online.
If one RAID1 member is fine, then the file system should be accessible.
It will just operate in degraded mode in the meantime.
2. what is the (S) mean ?
This is a hot spare, not sure why it isn't rebuilding now.... whilst
md0 is rebuilding.
3. how do i run fsck or other utility on ext4 filesystem ?
You only run fsck to fix a file system, not to fix a RAID set. If the
file system won't mount, then you could try fsck on the md0 or md1
device as required _before_ you mount it
4. do i run it on the raid1 device /dev/md0 or on the individual devices
/dev/sda and /devsdb
You would run fsck on the md0 device in your case, but that is your root
file system -- is it mounted now? If it is, then no fsck....
Your md1 seems to have been your swap partition, swap doesn't have any
file system on it to fsck. But it is now /mnt/raid ????
It might be useful for to supply output from blkid if you need more
help. And also output from the following:
mdadm --detail /dev/md0
mdadm --detail /dev/md1
That [mdadm --detail .... ] will give more details than /proc/mdstat
--
Kind Regards
AndrewM
Andrew McGlashan
Reply to: