question about raid 1

h725 · 09-15-2009, 07:31 AM

Hi,

I've a server with 2 IDE disk (80GB), with 3 partitions:

/boot (md0)
/ (md1)
swap (md2)

The second ide disk, hdb, have some hardware problem, so I'm going to replace it.
Meanwhile, I want to set hdb "fault" and remove it from raid.

Quote:

# mdadm /dev/md1 -f /dev/hdb2
mdadm: set /dev/hdb2 faulty in /dev/md1

Quote:

# mdadm /dev/md1 -r /dev/hdb2
mdadm: hot remove failed for /dev/hdb2: Device or resource busy

I think this is normal because / partition is mounted.. so how can I remove the device from raid? It's necessary to reboot system? In that case, I must edit /etc/mdadm/mdadm.conf before?

Another question:

Quote:

# mdadm --detail /dev/md1
/dev/md1:
Version : 00.90.01
Creation Time : Thu Feb 17 18:29:44 2005
Raid Level : raid1
Array Size : 79101696 (75.44 GiB 81.00 GB)
Device Size : 79101696 (75.44 GiB 81.00 GB)
Raid Devices : 2
Total Devices : 2
Preferred Minor : 1
Persistence : Superblock is persistent

Update Time : Tue Sep 15 10:33:58 2009
State : clean, degraded, recovering
Active Devices : 1
Working Devices : 2
Failed Devices : 0
Spare Devices : 1

Rebuild Status : 43% complete

UUID : fdb1286b:099c4e58aaf5:a35fd3824e86
Events : 0.76951239

Number Major Minor RaidDevice State
2 3 2 0 spare rebuilding /dev/hda2
1 3 66 1 active sync /dev/hdb2

Why there is one spare device?? I don't have any spare devices.. only the two disc.

Thankyou

ZenDrum · 09-15-2009, 02:29 PM

With most hardware IDE RAID controllers, for a RAID 1 failure, you shutdown the machine, swap the drive, and restart. When the controller sees the new drive, it will begin to rebuild the array.

Note this is for IDE, which normally doesn't support hot swap. If it were SAS, SATA, or some SCSI backplanes, you could hotswap and leave the machine running.

Hope this helps.

leandean · 09-15-2009, 11:06 PM

You have to remove the drive from the other arrays before you can remove it.

h725 · 09-16-2009, 07:11 AM

Quote:

Originally Posted by leandean

You have to remove the drive from the other arrays before you can remove it.

Yes, I've removed /dev/hdb1 and /dev/hdb3 from other arrays .. but /dev/hdb2 still busy:

Quote:

# mdadm /dev/md2 -f /dev/hdb3
mdadm: set /dev/hdb3 faulty in /dev/md2

# mdadm /dev/md2 -r /dev/hdb3
mdadm: hot removed /dev/hdb3

# mdadm /dev/md0 -f /dev/hdb1
mdadm: set /dev/hdb1 faulty in /dev/md0

# mdadm /dev/md0 -r /dev/hdb1
mdadm: hot removed /dev/hdb1

Quote:

# cat /proc/mdstat
Personalities : [linear] [raid1]
md1 : active raid1 hdb2[1] hda2[2]
79101696 blocks [2/1] [_U]
[===============>.....] recovery = 79.8% (63132672/79101696) finish=9.8min speed=26990K/sec
md2 : active raid1 hda3[0]
843584 blocks [2/1] [U_]

md0 : active raid1 hda1[0]
97664 blocks [2/1] [U_]

unused devices: <none>

Quote:

# mdadm /dev/md1 -r /dev/hdb2
mdadm: hot remove failed for /dev/hdb2: Device or resource busy

So it's necessary to reboot the machine? How can I tell md to not rebuild array /dev/md1 after the reboot?

Thankyou

DrLove73 · 09-16-2009, 02:20 PM

There is nice howto for removing failed PRIMARY HDD's from RAID1: http://www200.pair.com/mecham/raid/raid1-page3.html

However, is it possible that you are trying to remove the wrong HDD? You have [_U] for md1, but [U_] for md0 and md2. Be very carefull, check 3 times.

Also, from your posts looks like system is (automaticaly?) recovering failed hdX2 id md1 after you set the --failed flag.Notice this:

Quote:

Update Time : Tue Sep 15 10:33:58 2009
State : clean, degraded, recovering
Active Devices : 1
Working Devices : 2
Failed Devices : 0
Spare Devices : 1

Rebuild Status : 43% complete

even in your first post.

chrism01 · 09-17-2009, 01:26 AM

Check /etc/mdadm.conf & /etc/fstab and comment out refs to that raid.