[SOLVED]Problems with tuning my Raid

hawaix · 28. März 2013

Hello , thanks for this tuning .

This is mi Raid
Number Major Minor RaidDevice State
0 8 0 0 active sync /dev/sda
1 8 32 1 active sync /dev/sdc
2 8 48 2 active sync /dev/sdd
3 8 16 3 active sync /dev/sdb

I log in as root in ssh
sending this command

Code

mdadm --create /dev/md0 -e 1.2 -n 4 -c 128k -l 5 -N StorageRAID5 /dev/sda /dev/sdc /dev/sdd missing

I get an error
mdadm: invalid chunk/rounding value: 128k

help me THX

SerErris · 1. April 2013

Does your raid already exist before you try the mdadm command?

Your Raid needs to get destroyed before WARNING: ALL DATA WILL BE LOST!!!!

You need to remove all raid information on the discs with the mdadm --zero-superblock /dev/sda command for every disk.

BEWARE YOU WILL LOOSE ALL DATA ON THAT DISKS IF YOU CARRY OUT THE ABOVE COMMAND!!!

hawaix · 1. April 2013

Zitat

Does your raid already exist before you try the mdadm command?

Yes my raid already exist .

Zitat

BEWARE YOU WILL LOOSE ALL DATA ON THAT DISKS IF YOU CARRY OUT THE ABOVE COMMAND!!!

Thanks for info , I can get the command to destroy my raid?

hawaix · 1. April 2013

ok I destroyed my raid ,

Code

mdadm /dev/md127 --fail /dev/sda --remove /dev/sda

Code

mdadm /dev/md127 --fail /dev/sdb --remove /dev/sdb

Code

mdadm /dev/md127 --fail /dev/sdc --remove /dev/sdc

Code

mdadm /dev/md127 --fail /dev/sdd --remove /dev/sdd

Now what I do?

hawaix · 1. April 2013

before following this guide , I have to create the raid5? :roll:

SerErris · 1. April 2013

Now run this command:

Code

mdadm --create /dev/md0 -e 1.2 -n 4 -c 128k -l 5 -N StorageRAID5 /dev/sda /dev/sdb /dev/sdc /dev/sdd

This IS the command to create the RAID. It will also directly kick of the resync. After that you can create LVM or your filesystem however you like it.

hawaix · 2. April 2013

not work

This is my current raid5

I have not yet created the file system

I created the raid via Web Gui , is possible the guide step by step , to follow this guide the raid must be degraded mode ?

Many Thanks

SerErris · 2. April 2013

That is exactly what you wanted to achiev ... The raid has already been fully constructed. If you used the command above, it will have the right parameters set.

You can check with:

Code

mdadm --misc --detail /dev/md127

If everything is right, you should find the line
Chunk Size : 128K
like in this output:

Code

root@openmediavault:~# mdadm --misc --detail /dev/md127
/dev/md127:
        Version : 1.2
  Creation Time : Tue Oct  9 22:40:19 2012
     Raid Level : raid5
     Array Size : 5860540032 (5589.05 GiB 6001.19 GB)
  Used Dev Size : 1953513344 (1863.02 GiB 2000.40 GB)
   Raid Devices : 4
  Total Devices : 4
    Persistence : Superblock is persistent


    Update Time : Tue Apr  2 16:12:42 2013
          State : clean
 Active Devices : 4
Working Devices : 4
 Failed Devices : 0
  Spare Devices : 0


         Layout : left-symmetric
     Chunk Size : 128K


           Name : openmediavault:storage  (local to host openmediavault)
           UUID : 6a35a589:d9c8a30b:a61a95bf:c29f582b
         Events : 3494


    Number   Major   Minor   RaidDevice State
       0       8        0        0      active sync   /dev/sda
       1       8       16        1      active sync   /dev/sdb
       2       8       32        2      active sync   /dev/sdc
       4       8       48        3      active sync   /dev/sdd
root@openmediavault:~#

Alles anzeigen

So now skip the mdadm --add command and continue with the tuning

hawaix · 2. April 2013

Code

mdadm --misc --detail /dev/md0
/dev/md0:
        Version : 1.2
  Creation Time : Mon Apr  1 21:12:48 2013
     Raid Level : raid5
     Array Size : 8790795264 (8383.56 GiB 9001.77 GB)
  Used Dev Size : 2930265088 (2794.52 GiB 3000.59 GB)
   Raid Devices : 4
  Total Devices : 4
    Persistence : Superblock is persistent


    Update Time : Tue Apr  2 07:22:18 2013
          State : active
 Active Devices : 4
Working Devices : 4
 Failed Devices : 0
  Spare Devices : 0


         Layout : left-symmetric
     Chunk Size : 512K


           Name : Kymera:Storage  (local to host Kymera)
           UUID : 2e125ee9:7f116e9f:46fc7104:2a7bd088
         Events : 35


    Number   Major   Minor   RaidDevice State
       0       8        0        0      active sync   /dev/sda
       1       8       16        1      active sync   /dev/sdb
       2       8       32        2      active sync   /dev/sdc
       3       8       48        3      active sync   /dev/sdd

Alles anzeigen

My Chunk Size is 512K .

I created the raid 5 via webgui , because when I give the command

Code

root@Kymera:~# mdadm --create /dev/md0 -e 1.2 -n 4 -c 128k -l 5 -N Storage /dev/sda /dev/sdb /dev/sdc /dev/sdd
mdadm: invalid chunk/rounding value: 128k

SerErris · 2. April 2013

Ah okay, there is a typo in it... sorry for that.

Please try:

Code

mdadm --create /dev/md0 -e 1.2 -n 4 -c 128 -l 5 -N Storage /dev/sda /dev/sdb /dev/sdc /dev/sdd

I am not sure why I put the k in, but you are right, it does not work with k ... so simply leave it away as -c expects a k value.

Can you give it a try with that? I then will update the original post to reflect that.

hawaix · 2. April 2013

Thanks SerErris, without k

Code

mdadm --create /dev/md0 -e 1.2 -n 4 -c 128 -l 5 -N Storage /dev/sda /dev/sdb /dev/sdc /dev/sdd

I get this error
mdadm: super1.x cannot open /dev/sda: Device or resource busy
mdadm: /dev/sda is not suitable for this array.
mdadm: super1.x cannot open /dev/sdb: Device or resource busy
mdadm: /dev/sdb is not suitable for this array.
mdadm: super1.x cannot open /dev/sdc: Device or resource busy
mdadm: /dev/sdc is not suitable for this array.
mdadm: super1.x cannot open /dev/sdd: Device or resource busy
mdadm: /dev/sdd is not suitable for this array.
mdadm: create aborted

My Raid5 is currently active (see my screenshot)

SerErris · 2. April 2013

yes of cause ... you need to destroy the raid before (use the gui for that).

hawaix · 2. April 2013

Zitat von "SerErris"

yes of cause ... you need to destroy the raid before (use the gui for that).

Ok i destroy my Raid , and after

Code

mdadm --create /dev/md0 -e 1.2 -n 4 -c 128 -l 5 -N Storage /dev/sda /dev/sdb /dev/sdc /dev/sdd

mdadm: partition table exists on /dev/sda but will be lost or
meaningless after creating array
mdadm: partition table exists on /dev/sdb but will be lost or
meaningless after creating array
mdadm: partition table exists on /dev/sdc but will be lost or
meaningless after creating array
mdadm: partition table exists on /dev/sdd but will be lost or
meaningless after creating array
Continue creating array?
Continue creating array? (y/n) y
mdadm: array /dev/md0 started.

Code

Version : 1.2
  Creation Time : Tue Apr  2 17:21:26 2013
     Raid Level : raid5
     Array Size : 8790796032 (8383.56 GiB 9001.78 GB)
  Used Dev Size : 2930265344 (2794.52 GiB 3000.59 GB)
   Raid Devices : 4
  Total Devices : 4
    Persistence : Superblock is persistent


    Update Time : Tue Apr  2 17:25:21 2013
          State : clean, degraded, recovering
 Active Devices : 3
Working Devices : 4
 Failed Devices : 0
  Spare Devices : 1


         Layout : left-symmetric
     Chunk Size : 128K


 Rebuild Status : 0% complete


           Name : Kymera:Storage  (local to host Kymera)
           UUID : 51ff4b7a:919db2bc:68ed45fd:ae8accda
         Events : 114


    Number   Major   Minor   RaidDevice State
       0       8        0        0      active sync   /dev/sda
       1       8       16        1      active sync   /dev/sdb
       2       8       32        2      active sync   /dev/sdc
       4       8       48        3      spare rebuilding   /dev/sdd

Alles anzeigen

I can now create the file system?

hawaix · 2. April 2013

Rebuild Status : 2% complete , omg maybe takes 10 hours

SerErris · 2. April 2013

yes, you can now create the file system.

And as I see now, you have partitions on it. It wont hurt, but may confuse that there is a partition and you do not use it. You should not use any partition that is okay, but it is now looking weird.

However no you can continue with the filesystem. You should not reboot until the reconstruct is done, otherwise it will start again.

hawaix · 3. April 2013

Zitat

yes, you can now create the file system.

Yes i created the filesystem .

Code

Version : 1.2
  Creation Time : Tue Apr  2 17:21:26 2013
     Raid Level : raid5
     Array Size : 8790796032 (8383.56 GiB 9001.78 GB)
  Used Dev Size : 2930265344 (2794.52 GiB 3000.59 GB)
   Raid Devices : 4
  Total Devices : 4
    Persistence : Superblock is persistent


    Update Time : Wed Apr  3 05:01:49 2013
          State : clean
 Active Devices : 4
Working Devices : 4
 Failed Devices : 0
  Spare Devices : 0


         Layout : left-symmetric
     Chunk Size : 128K


           Name : Kymera:Storage  (local to host Kymera)
           UUID : 51ff4b7a:919db2bc:68ed45fd:ae8accda
         Events : 2898


    Number   Major   Minor   RaidDevice State
       0       8        0        0      active sync   /dev/sda
       1       8       16        1      active sync   /dev/sdb
       2       8       32        2      active sync   /dev/sdc
       4       8       48        3      active sync   /dev/sdd

Alles anzeigen

Next step

Code

tune2fs -E stride=32,stripe-width=96 -Odir_index /dev/mapper/MYUUID

(what is this ? Odir_index)
I get this error

Code

Can not find a valid superblock for the filesystem.

Any idea?

SerErris · 3. April 2013

I am not sure if a space is an issue here.

Try

Code

tune2fs -E stride=32,stripe-width=96 -O dir_index /dev/mapper/MYUUID

And UUID is NOT the one from the RAID (51ff4b7a:919db2bc:68ed45fd:ae8accda) you need the on from your ext4 filesystem.

Another question, do you use LVM? If not, the /dev/mapper thing is wrong and you need to find the UUID of the ext4 of your mount device, which would be /dev/mdX.

Those are the few things that come to my mind.

hawaix · 3. April 2013

Zitat

Another question, do you use LVM?

No i do not use LVM

I found the UUID from my ext4

Zitat

/dev/md0: LABEL="StorageRAID5" UUID="fb054e34-b5ed-4f03-b44c-92bb4b507973" TYPE="ext4"

Code

tune2fs -E stride=32,stripe-width=96 -O dir_index /dev/mapper/fb054e34-b5ed-4f03-b44c-92bb4b507973

i get this error

Code

tune2fs 1.41.12 (17-May-2010)

Code

No such file or directory when opening /dev/mapper/fb054e34-b5ed-4f03-b44c-92bb4b507973

Code

Can not find a valid superblock for the filesystem

my question is I have to continue tuning or I can stop ?if I stop there may be problems of data loss?

br41n · 3. April 2013

you should use the command

Code

tune2fs -E stride=32,stripe-width=96 -O dir_index /dev/mdX

X is for your mdadm raid, in your case it is 0 in some cases it is 127

SerErris · 3. April 2013

Yep, the whole thread is using RAID+LVM+Ext4 ... so if you leave LVM away, you then need to use the right device for the tune2fs command as br41n stated.

Jetzt mitmachen!