Nigel Kendrick spake the following on 10/5/2006 3:48 PM:

-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Scott Silva Sent: Thursday, October 05, 2006 11:15 PM To: centos@centos.org Subject: [CentOS] Re: Cannot re-make a software raid pair

Nigel Kendrick spake the following on 10/5/2006 12:45 PM:

...

Apologies if you get this twice - the first one didn't seem to make it...

Hi Guys,

I have just replaced a faulty Max...woah, wait...this one's a Seagate... IDE hard disk but I cannot remake the software raid pair.

The currently running disk is hda and I am trying to add back hdg - both are master drives on separate controllers.

I have run fdisk on hdg and created the same partition structure as had and changed their types to 'fd' but when I try and rebuild the array I

get:

...

mdadm /dev/md0 -a /dev/hdg1

mdadm: hot add failed for /dev/hdg1: Invalid argument

If I look at /var/log/messages, I have:

Oct 5 16:04:39 petdoctors kernel: md: could not bd_claim hdg1. Oct 5 16:04:39 petdoctors kernel: md: error, md_import_device() returned -16

...

From other postings, this seems to imply the hdg1 partition is mounted but I

can find no evidence of this (df or cat /proc/mounts or lsof). I have done sone googling and found others having fun with this but no specific solution.

Anyone?

Maybe the table didn't get re-read after fdisk was done. I usually use sfdisk just because I can copy a partition table without making my usual fat-fingered typos. Are you sure that the new drive isn't smaller than the old? Even by a few cylinders.

--

Hi Scott,

Disks are 100% identical - right down to the last cylinder...hda is the live disk and hdg is the new one...

[root@petdoctors ~]# sfdisk --no-reread /dev/hda

Disk /dev/hda: 24321 cylinders, 255 heads, 63 sectors/track Old situation: Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0

Device Boot Start End #cyls #blocks Id System /dev/hda1 * 0+ 12 13- 104391 fd Linux raid autodetect /dev/hda2 13 273 261 2096482+ 82 Linux swap /dev/hda3 274 24320 24047 193157527+ fd Linux raid autodetect /dev/hda4 0 - 0 0 0 Empty Input in the following format; absent fields get a default value. <start> <size> <type [E,S,L,X,hex]> <bootable [-,*]> <c,h,s> <c,h,s> Usually you only need to specify <start> and <size> (and perhaps <type>).

/dev/hda1 : [root@petdoctors ~]# sfdisk /dev/hdg Checking that no-one is using this disk right now ... OK

Disk /dev/hdg: 24321 cylinders, 255 heads, 63 sectors/track Old situation: Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0

Device Boot Start End #cyls #blocks Id System /dev/hdg1 0+ 12 13- 104391 fd Linux raid autodetect /dev/hdg2 13 273 261 2096482+ 82 Linux swap /dev/hdg3 274 24320 24047 193157527+ fd Linux raid autodetect /dev/hdg4 0 - 0 0 0 Empty Input in the following format; absent fields get a default value. <start> <size> <type [E,S,L,X,hex]> <bootable [-,*]> <c,h,s> <c,h,s> Usually you only need to specify <start> and <size> (and perhaps <type>).

/dev/hdg1 :

[root@petdoctors ~]# mdadm /dev/md1 -a /dev/hdg3 mdadm: hot add failed for /dev/hdg3: Invalid argument [root@petdoctors ~]#

I know this is a real basic question, but did you fail then remove the old drive first?

Nigel Kendrick spake the following on 10/6/2006 3:41 AM:

-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Scott Silva Sent: Thursday, October 05, 2006 11:57 PM To: centos@centos.org Subject: [CentOS] Re: Cannot re-make a software raid pair

Nigel Kendrick spake the following on 10/5/2006 3:48 PM:

...

-----Original Message----- From: centos-bounces@centos.org [mailto:centos-bounces@centos.org] On Behalf Of Scott Silva Sent: Thursday, October 05, 2006 11:15 PM To: centos@centos.org Subject: [CentOS] Re: Cannot re-make a software raid pair

Nigel Kendrick spake the following on 10/5/2006 12:45 PM:

...

Apologies if you get this twice - the first one didn't seem to make it...

...

...

Hi Guys,

I have just replaced a faulty Max...woah, wait...this one's a Seagate... IDE hard disk but I cannot remake the software raid pair.

[SNIP]

I know this is a real basic question, but did you fail then remove the old drive first?

Hi Scott,

Yes I did, but half the problem is/may be that the faulty drive is not even 'seen' by the server and so software RAID takes it out of the mirrored pairing on boot and I am pretty sure that when I 'failed' it it threw an error that the drive wasn't there anyway. I will try again and also see if I can get the drive recognised so I can 'fail' it.

Thanks again for the feedback

Nigel

I think you can add a--force to the remove command. Otherwise the drive is still included in the array in the metadata on the good drive.

On 05/10/06, Nigel Kendrick support-lists@petdoctors.co.uk wrote:

[root@petdoctors ~]# mdadm /dev/md1 -a /dev/hdg3 mdadm: hot add failed for /dev/hdg3: Invalid argument [root@petdoctors ~]#

I don't know md at all well, but from a quick RTFM, has the previous, faulty /dev/hdg3 been removed from the software RAID device?

http://unthought.net/Software-RAID.HOWTO/Software-RAID.HOWTO-6.html#ss6.3

<------------------------------------------------------------------------> Users of mdadm can run the command

# mdadm --detail /dev/md1

Now you've seen how it goes when a device fails. Let's fix things up. First, we will remove the failed disk from the array. Users of mdadm can run the command

# mdadm /dev/md1 -r /dev/sdc2

Now we have a /dev/md1 which has just lost a device. This could be a degraded RAID or perhaps a system in the middle of a reconstruction process. We wait until recovery ends before setting things back to normal.

So the trip ends when we send /dev/sdc2 back home. As usual, you can use mdadm instead of raidtools. This should be the command

# mdadm /dev/md1 -a /dev/sdc2

As the prodigal son returns to the array, we'll see it becoming an active member of /dev/md1 if necessary. If not, it will be marked as an spare disk. That's management made easy. <------------------------------------------------------------------------>

Will.