Re: Multipath problems with 2.6.8.1

["Jaime Peñalba" <jpenalba@corenetworks.es>]

Im using multipath for failover, not for round robin (rr over multipath 
isnt supported with the latest stable kernel), also i was wondering 
about how this should be configured before sending a mail to the list, 
so i  tested both ways (as "2 raid disks" and as "raid disk with spare 
disk"), the good way is using a spare disk, or i think so. On raidtools2 
there is an example for multipath and they use a spare disk, and it has 
sense, because it should act as a spare disk like on a raid 1/5, it uses 
the raid disk path, and if it fails, it uses the spare disk path.

Anyway here is the output doing it as 2 raid disks

1) create md
apache2:~# mdadm --create /dev/md0 --force --level multipath 
--raid-devices 2 /dev/sdb1 /dev/sdd1
mdadm: array /dev/md0 started.

apache2:~# cat /proc/mdstat
Personalities : [multipath]
md0 : active multipath sdd1[1] sdb1[0]
      10485632 blocks [2/2] [UU]
     
unused devices: <none>

---------

The output from dmesg

md: bind<sdb1>
md: bind<sdd1>
multipath: array md0 active with 2 out of 2 IO paths


And it should be something like this (configured with spare)

md: bind<sdb1>
md: bind<sdd1>
multipath: array md0 active with 1 out of 1 IO paths
MULTIPATH conf printout:
--- wd:1 rd:1
disk0, o:1, dev:sdb1
MULTIPATH conf printout:
--- wd:1 rd:1
disk0, o:1, dev:sdb1


2) stop md
apache2:~# mdadm -S /dev/md0


3) start md again
apache2:~# mdadm --assemble /dev/md0 /dev/sdb1 /dev/sdd
mdadm: failed to add /dev/sdb1 to /dev/md0: Device or resource busy
mdadm: /dev/md0 has been started with 2 drives and -1 spares.

apache2:~# cat /proc/mdstat
Personalities : [multipath]
md0 : active multipath sdd1[0]
      10485632 blocks [1/1] [U]
     
unused devices: <none>

-------------------------------

Superblocks info

apache2:~# mdadm -E /dev/sdb1
/dev/sdb1:
          Magic : a92b4efc
        Version : 00.90.00
           UUID : 57d730b7:49b3d124:b5d30806:127c5bc3
  Creation Time : Thu Sep  9 19:49:10 2004
     Raid Level : multipath
    Device Size : 10485632 (10.00 GiB 10.74 GB)
   Raid Devices : 2
  Total Devices : 1
Preferred Minor : 0

    Update Time : Thu Sep  9 20:16:37 2004
          State : dirty
 Active Devices : 1
Working Devices : 1
 Failed Devices : 1
  Spare Devices : 0
       Checksum : 9626eafd - correct
         Events : 0.45


      Number   Major   Minor   RaidDevice State
this     0       8       49        0      active sync   /dev/sdd1
   0     0       8       49        0      active sync   /dev/sdd1
   1     1       0        0        1      faulty removed

-----------------------

apache2:~# mdadm -E /dev/sdd1
/dev/sdd1:
          Magic : a92b4efc
        Version : 00.90.00
           UUID : 57d730b7:49b3d124:b5d30806:127c5bc3
  Creation Time : Thu Sep  9 19:49:10 2004
     Raid Level : multipath
    Device Size : 10485632 (10.00 GiB 10.74 GB)
   Raid Devices : 2
  Total Devices : 1
Preferred Minor : 0

    Update Time : Thu Sep  9 20:16:37 2004
          State : dirty
 Active Devices : 1
Working Devices : 1
 Failed Devices : 1
  Spare Devices : 0
       Checksum : 9626eafd - correct
         Events : 0.45


      Number   Major   Minor   RaidDevice State
this     0       8       49        0      active sync   /dev/sdd1
   0     0       8       49        0      active sync   /dev/sdd1
   1     1       0        0        1      faulty removed


As you can see both paths points to the same disk, because sdb1 has the 
sdd1 superblock, that is the last superblock wroten.


Jaime.


Guy wrote:

>Somehow you confused it!  And me.
>
>You used multipath and spare, but only listed 2 devices.
>I don't think a multipath array can have a spare disk.
>
>Are you sure the 2 paths point to the same disk?
>
>I have never user multipath, so I am just guessing.
>
>If you have no data you need to save yet, try this:
>mdadm --create /dev/md0 --force --level multipath /dev/sdb1 /dev/sdd1
>or
>mdadm --create /dev/md0 --force --level multipath --raid-devices=2 /dev/sdb1
>/dev/sdd1
>
>Guy
>
>-----Original Message-----
>From: linux-raid-owner@vger.kernel.org
>[mailto:linux-raid-owner@vger.kernel.org] On Behalf Of Jaime Peñalba
>Sent: Thursday, September 09, 2004 11:33 AM
>To: linux-raid@vger.kernel.org
>Cc: neilb@cse.unsw.edu.au
>Subject: Multipath problems with 2.6.8.1
>
>Hi,
>
>Im having some troubles with multipath on a 2.6.8.1 linux kernel.
>
>Im using 2 QLogic 2344 connected to the SAN, i think that my problem is 
>related to superblocks, because when i create a new multipath device, it 
>writes the superblock for both disks and really is the same disk trought 
>two paths. I have else tried to disable persistent superblock, but mdadm 
>doesnt support multipath with the build mode.
>
>Im doing it like this
>
>1) Start the md
>apache1:~/soft/mdadm-1.7.0# mdadm --create /dev/md0 --force --level 
>multipath --raid-devices=1 /dev/sdb1 --spare-devices=1 /dev/sdd1
>mdadm: array /dev/md0 started.
>
>apache1:~/soft/mdadm-1.7.0# cat /proc/mdstat
>Personalities : [multipath]
>md0 : active multipath sdd1[1] sdb1[0]
>      4882304 blocks [1/1] [U]
>     
>unused devices: <none>
>
>This is ok and runs fine, But superblock is only at /dev/sdd1, so no 
>remount would be possible, or i think so...
>
>apache1:~/soft/mdadm-1.7.0# mdadm -E /dev/sdb1
>mdadm: No super block found on /dev/sdb1 (Expected magic a92b4efc, got 
>00000000)
>
>apache1:~/soft/mdadm-1.7.0# mdadm -E /dev/sdd1
>/dev/sdd1:
>          Magic : a92b4efc
>        Version : 00.90.00
>           UUID : af3395fd:fac2b820:2ca7083c:612864e0
>  Creation Time : Thu Sep  9 17:20:43 2004
>     Raid Level : multipath
>    Device Size : 4882304 (4.66 GiB 5.00 GB)
>   Raid Devices : 1
>  Total Devices : 2
>Preferred Minor : 0
>
>    Update Time : Thu Sep  9 17:20:43 2004
>          State : dirty
> Active Devices : 1
>Working Devices : 2
> Failed Devices : 0
>  Spare Devices : 1
>       Checksum : 63bc7489 - correct
>         Events : 0.36
>
>
>      Number   Major   Minor   RaidDevice State
>this     1       8       49        1      spare   /dev/sdd1
>   0     0       8       17        0      active sync   /dev/sdb1
>   1     1       8       49        1      spare   /dev/sdd1
>
>
>
>2) Stop the md
>apache1:~/soft/mdadm-1.7.0# mdadm -S /dev/md0
>
>
>3) Start the md again
>apache1:~/soft/mdadm-1.7.0# mdadm --assemble /dev/md0 
>--uuid=af3395fd:fac2b820:2ca7083c:612864e0 /dev/sdb1 /dev/sdd1
>mdadm: no RAID superblock on /dev/sdb1
>mdadm: /dev/sdb1 has wrong uuid.
>mdadm: /dev/md0 has been started with 1 drive.
>
>
>apache1:~/soft/mdadm-1.7.0# cat /proc/mdstat
>Personalities : [multipath]
>md0 : active multipath sdd1[0]
>      4882304 blocks [1/1] [U]
>     
>unused devices: <none>
>
>Now things arent working very well....
>
>apache1:~/soft/mdadm-1.7.0# mdadm --detail /dev/md0
>/dev/md0:
>        Version : 00.90.01
>  Creation Time : Thu Sep  9 17:20:43 2004
>     Raid Level : multipath
>     Array Size : 4882304 (4.66 GiB 5.00 GB)
>   Raid Devices : 1
>  Total Devices : 1
>Preferred Minor : 0
>    Persistence : Superblock is persistent
>
>    Update Time : Thu Sep  9 17:26:21 2004
>          State : clean
> Active Devices : 1
>Working Devices : 1
> Failed Devices : 0
>  Spare Devices : 0
>
>
>    Number   Major   Minor   RaidDevice State
>       0       8       49        0      active sync   /dev/sdd1
>           UUID : af3395fd:fac2b820:2ca7083c:612864e0
>         Events : 0.37
>
>
>I would appreciate very much any help with this.
>
>Thanks,
>Jaime.
>
>  
>

-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html