Can't access drive - Need raid5/linux help

[Jocko]

Sorry , to have been lazy to detail how to mount your raid...

So If I understand you keep open the telnet window since few days.

- First install the mount binary as detailed in viewtopic.php?f=26&t=1574&start=40#p19163

- create a mount point

Code: Select all

mkdir /md4

- try to mount md4

Code: Select all

/usr/sbin/mount /dev/md4 /md4

If you succeeded, you can see your data with the command

Code: Select all

ls -l /md4

[imt17]

- try to mount md4

Code: Select all

/usr/sbin/mount /dev/md4 /md4

After I ran the above command I got the following:
mount: /dev/md4: can't read superblock

I have my machine at home connected via telnet ever since I got it connected last night. I am currently connected via teamviewer from my office to this win XP at home, so I can run the stuff here, when I get responses, since I am not sure if we are all in the same time zone :mrgreen:

[Jocko]

So fs on your raid is corrupted and you need to repair it.

read the other posts (still in the same topic viewtopic.php?f=26&t=1574&start=40#p19163) about xfs_repair.

[imt17]

So fs on your raid is corrupted and you need to repair it.

read the other posts (still in the same topic viewtopic.php?f=26&t=1574&start=40#p19163) about xfs_repair.

No fs on your Lacie NAs is xfs and not at all ext3.
How do I know for sure that its xfs and not ext3, etc? That question came up in that thread as well. I didn;t see where/how to tell

I don't see how to reference a particular post but on page 6 of that the thread you referenced. But, second to last post on that page, fvdw posted:

--edit see here how to use xfs_repair xfs-repair info
Running it with option -n will do only a check.

Ps you will need to install also the package glibc-mini-mkfs.xfs-25feb14 to be able to use xfs-repair
Uploud the tar archive in the glibc-mini...zip file to the 5big2 via telnet terminial and extract it.
Then after that upload the file xfs-repair in the xfs_repair.zip file and place it folder /usr/sbin and make it executable.
Now you should be able to use xfs-repair

In order to upload these files do I do?

Code: Select all

tftp -l glibc-mini-mkfs.xfs-25feb14.tar -r glibc-mini-mkfs.xfs-25feb14.tar -g ip-pc
tftp -l xfs_repair.tar -r xfs_repair.tar -g ip-pc

These files are already included with the fvdw-sl console version 5.5.

Do I need to do anything first with the drives before this? Dismount or anything? Don't want to mess this up.
No do nothing

Also, is there a reason not to first try (saw this online with something else):

Code: Select all

mdadm --stop /dev/md4

if you stop the raid you can not check its fs later!
Then run the command you gave me before:

Code: Select all

mdadm --assemble --force /dev/md0 /dev/sda2 /dev/sdb2 /dev/sdc2 missing /dev/sde2

To see if it builds the raid5 array using the 4 out of 5 disks instead of what appears to now be only 3 out of 5 disks that currently make up md4, that I currently had assembled? Unless I am reading it wrong.
I made an error previously: missing option can be used only when your create a raid. So the right command is

Code: Select all

mdadm --assemble --force /dev/md0 /dev/sda2 /dev/sdb2 /dev/sdc2 /dev/sde2

Please not to try several times to disassemble/assemble your raid :disapprove

[Jocko]

I edited your last post.

[imt17]

Thanks.

Confused on two parts.

1) I know the files for xfs are included in 5.5 but don;t they have to be uploaded from the win pc to the Lacie to actually use them? I had to do that with mdadm. If so is the command(s) I wrote to upload those two files correct or something else?

Then once I upload what command(s) should I run and what should I post here.

2) I see you corrected the assemble syntax. However, should I run that first or not since my raid is already assembled the way it is and you want me to leave it alone?

[Jocko]

To install xps_repair after uploading, do

Code: Select all

 tar -xf glibc-mini-mkfs.xfs-25feb14.tar -C /

and

Code: Select all

 tar -xf xfs_repair.tar -C /

and try first a safe check

Code: Select all

xfs_repair -n /dev/md4

(no modify mode)

2) I see you corrected the assemble syntax. However, should I run that first or no since my raid is already assembled the way it is and you want me to leave it alone the way it is?

I am sorry but I am lost...about where you are.

I repeat if the raid is already assembled no need to try again

So post again

Code: Select all

cat /proc/mdstat

and also
 mdadm --detail /dev/md4


[imt17]

Code: Select all

 tar -xf xfs_repair.tar -C /

I got the following error:
tar: invalid tar magic

So post again

Code: Select all

cat /proc/mdstat

and also
 mdadm --detail /dev/md4


Here you go

root@fvdw-sta-kirkwood:/ # cat /proc/mdstat
Personalities : [linear] [raid0] [raid1] [raid10] [raid6] [raid5] [raid4]
md4 : inactive sda2[0](S) sde2[4](S) sdc2[2](S) sdb2[1](S)
3898949796 blocks super 1.0


mdadm: md device /dev/md4 does not appear to be active.

[Jocko]

Before using xfs_repair, seems that there is something to deepen on the raid side.

Maybe as you tried to build the raid several times, the devices have now a spare status (S) => no fs is currently available on the raid!
So do

Code: Select all

mdadm --stop /dev/md4
mdadm --assemble /dev/md4 --force --run --uuid=36d39884:a2d4762b:3e519be3:fe529c84

and post again

Code: Select all

cat /proc/mdstat
 mdadm --detail /dev/md4

[imt17]

Before using xfs_repair, seems that there is something to deepen on the raid side.

Maybe as you tried to build the raid several times, the devices have now a spare status (S) => no fs is currently available on the raid!

I went back to the start to view on the telnet window. It always was this was with the S. raid was always inactive.

Anyway after I ran the assemble command you posted I got the following error:mdadm: an md device must be given in this mode