Re: Mdadm drive fail power cord issue.

See in the above that md1 is active on /dev/sdc1 while the other

devices are all active in /dev/sdb{2,3,4}. This means you need to be
careful which disk you do things to since you could wipe out the other
disk if not careful. Three are one way and one is the other way.

hmm i have taken the backup on external USB however, just learning my challenge is not to loose the data :) becuase in other cases i may not found the option of USB backup.

> > You will need to know that information in order to proceed successfully.
> > I would like to see the output of:
> >
> > mdadm --examine /dev/sdc2
>
> root@nasbox:~# mdadm --examine /dev/sdc2
> mdadm: No md superblock detected on /dev/sdc2.

That was an example from my system. For your system you would use these:

mdadm --examine /dev/sdc1
mdadm --examine /dev/sdb2
mdadm --examine /dev/sdb3
mdadm --examine /dev/sdb4

Noting that the first is on sdc and the other three on sdb.

Thanks i got that!

> root@nasbox:~# sfdisk -d /dev/sdb
>
> WARNING: GPT (GUID Partition Table) detected on '/dev/sdb'! The util sfdisk
> doesn't support GPT. Use GNU Parted.

Hmm... I am not a GPT expert. I haven't been using that partition
table type. Yet. It is inevitably in the future. But that means my
question was poorly written. Sorry.

yes agree and sorry for that incomplete or less comprehensive question.

however the limit was 2TB HD and finally i had to use GPT but this time i baught 1.5TBx2 drives so this time i may not face the GBT difficulty and thing can be worked out with Fdisk alone.

> root@nasbox:~# sfdisk -d /dev/sdc
> read: Input/output error

That looks scary to me. You have an active partition on /dev/sdc1 but
here /dev/sdc reported an I/O error? That seems very bad. At this
point I would stop and look carefully at the drive. I would look at
the SMART selftests from the drive. Here are some useful commands.

as i said i messed up the first partition :( and loss all the data.

smartctl -i /dev/sdc
smartctl -l error /dev/sdc
smartctl -t short /dev/sdc
smartctl -l selftest

# smartctl -i /dev/sdc                                          smartctl 5.43 2012-05-01 r3539 [i686-linux-2.6.32-5-686] (local build)
Copyright (C) 2002-12 by Bruce Allen, http://smartmontools.sourceforge.net

Vendor:               /3:0:0:0
Product:
User Capacity:        600,332,565,813,390,450 bytes [600 PB]
Logical block size:   774843950 bytes
scsiModePageOffset: response length too short, resp_len=47 offset=50 bd_len=46
>> Terminate command early due to bad response to IEC mode page
A mandatory SMART command failed: exiting. To continue, add one or more '-T permissive' options.

root@nasbox:/dev# smartctl -l error /dev/sdc
smartctl 5.43 2012-05-01 r3539 [i686-linux-2.6.32-5-686] (local build)
Copyright (C) 2002-12 by Bruce Allen, http://smartmontools.sourceforge.net

Short INQUIRY response, skip product id
A mandatory SMART command failed: exiting. To continue, add one or more '-T permissive' options.

root@nasbox:/dev# smartctl -l selftest /dev/sdc
smartctl 5.43 2012-05-01 r3539 [i686-linux-2.6.32-5-686] (local build)
Copyright (C) 2002-12 by Bruce Allen, http://smartmontools.sourceforge.net

Short INQUIRY response, skip product id
A mandatory SMART command failed: exiting. To continue, add one or more '-T permissive' options.

But if you are getting I/O errors from /dev/sdc then that probably
explains why you were getting an invalid argument error from mdadm
talking to that disk.

> sfdisk: read error on /dev/sdc - cannot read sector 0
> /dev/sdc: unrecognized partition table type
> No partitions found

I don't know what to think. How can the array be active on /dev/sdc1
when /dev/sdc is giving I/O errors? Anyone else have any ideas? I
think the disk may be failing. Hopefully the SMART selftests will
confirm it.

I am going to ignore the I/O errors for a moment and continue as if
that didn't happen. But it did. So beware.

Normally I would say that you need to clone your partition from one
disk to the other. I would normally use 'sfdisk -d /dev/sdb | sfdisk
/dev/sdc' to do it. But you are using GPT tables so that won't work.
Plus you have an active array on /dev/sdc2 that you don't want to
disrupt. So instead I can only say that you need to make sure that
sdc2, sdc3, sdc4 is partitioned like sdb2, sdb3, sdb4 and that sdb1 is
partitioned like sdc1. Then add them correct member back into the
array. But I don't know how to clone GPT tables. Perhaps someone
else will have help for that part. And I don't understand the I/O
error and think that needs to be understood first.

i got your point and to decrease my paint of using GPT i have just baught 1.5 TB harddrives.

so i am creating another storage server. i have the backup and restoring all the data to new fileserver
but the problem part is samba :( actually all users UID and GID would be lost and i have to reassign

all the rights once again to all the folders that a paint.

do you or any one know, how can i map samba MAP correct UID and GID to new users as the old users were using in my old Samba box.

i know there are some command line games but assigning individually a UID or GID via command would be hell of a tough job.

Good luck!
Bob

Thanks for your help bob