[Ocfs2-users] Different disksizes on same disks between cluster nodes

[Herbert van den Bergh]

Before anyone else wastes any time looking into this "mystery": one df 
uses -h and the other uses -H

Thanks,
Herbert.


On 03/22/2011 01:13 PM, Sunil Mushran wrote:
> Can you run df without any arguments? Could be a rounding issue in df.
> Still unsure how ocfs2 install/deinstall could trigger that.
>
> On 03/22/2011 01:52 AM, Kristiansen Morten wrote:
>>
>> Hi,
>>
>> Resently we resinstalled our four node cluster and tried to upgrade
>> ocfs2 from 1.2.9 to 1.4.7. When two nodes was reinstalled, I stopped
>> ocfs2, deinstalled ocfs2 version 1.2.9 and installed 1.4.7. The last
>> two nodes was down for reinstall. When I mounted node 2 I discovered a
>> mismatch between disksizes on node 1 and node2:
>>
>> [root at node1 ocfs2]# df -Ph
>>
>> Filesystem Size Used Avail Use% Mounted on
>>
>> /dev/sda11 965M 623M 293M 69% /
>>
>> /dev/sda10 449G 848M 425G 1% /disk00
>>
>> /dev/sda9 2.0G 60M 1.8G 4% /tmp
>>
>> /dev/sda7 9.7G 151M 9.1G 2% /home
>>
>> /dev/sda6 9.7G 1.8G 7.5G 19% /usr
>>
>> /dev/sda5 9.7G 416M 8.8G 5% /var
>>
>> /dev/sda8 4.5G 171M 4.1G 4% /var/log
>>
>> /dev/sda3 20G 5.6G 13G 31% /opt/oracle
>>
>> /dev/sda1 190M 12M 169M 7% /boot
>>
>> tmpfs 16G 0 16G 0% /dev/shm
>>
>> /dev/mapper/mpath8 2.0T 1.9T 158G 93% /disk01
>>
>> /dev/mapper/mpath9 700G 26G 675G 4% /disk02
>>
>> /dev/mapper/mpath10 550G 162G 389G 30% /disk03
>>
>> /dev/mapper/mpath0 850G 652G 199G 77% /disk04
>>
>> /dev/mapper/mpath6 200G 18G 183G 9% /disk05
>>
>> /dev/mapper/mpath5 450G 240G 211G 54% /disk06
>>
>> /dev/mapper/mpath3 750G 2.6G 748G 1% /disk07
>>
>> /dev/mapper/mpath2 3.6T 2.2T 1.5T 60% /disk08
>>
>> /dev/mapper/mpath4 4.0G 392M 3.7G 10% /disk09
>>
>> /dev/mapper/mpath1 316G 5.1G 311G 2% /disk10
>>
>> /dev/mapper/mpath7 4.1G 392M 3.7G 10% /disk11
>>
>> [root at node2 ocfs2]# df -PH
>>
>> Filesystem Size Used Avail Use% Mounted on
>>
>> /dev/sda11 1.1G 654M 307M 69% /
>>
>> /dev/sda10 482G 890M 457G 1% /disk00
>>
>> /dev/sda9 2.1G 63M 2.0G 4% /tmp
>>
>> /dev/sda7 11G 159M 9.8G 2% /home
>>
>> /dev/sda6 11G 1.9G 8.0G 19% /usr
>>
>> /dev/sda5 11G 436M 9.5G 5% /var
>>
>> /dev/sda8 4.8G 175M 4.4G 4% /var/log
>>
>> /dev/sda3 21G 6.0G 14G 31% /opt/oracle
>>
>> /dev/sda1 200M 13M 177M 7% /boot
>>
>> tmpfs 17G 0 17G 0% /dev/shm
>>
>> /dev/mapper/mpath0 2.2T 2.0T 170G 93% /disk01
>>
>> /dev/mapper/mpath1 752G 28G 725G 4% /disk02
>>
>> /dev/mapper/mpath2 591G 173G 418G 30% /disk03
>>
>> /dev/mapper/mpath3 913G 700G 213G 77% /disk04
>>
>> /dev/mapper/mpath9 215G 19G 197G 9% /disk05
>>
>> /dev/mapper/mpath8 484G 257G 227G 54% /disk06
>>
>> /dev/mapper/mpath6 806G 2.8G 803G 1% /disk07
>>
>> /dev/mapper/mpath5 4.0T 2.4T 1.6T 60% /disk08
>>
>> /dev/mapper/mpath7 4.3G 411M 3.9G 10% /disk09
>>
>> /dev/mapper/mpath4 339G 5.5G 334G 2% /disk10
>>
>> /dev/mapper/mpath10 4.4G 411M 4.0G 10% /disk11
>>
>> When I discovered the mismatch, I stopped ocfs2, deinstalled version
>> 1.4.7 and reinstalled 1.2.9 on both nodes. When mounting the disks
>> again, the sizes was the same and the same as node1 when running
>> 1.4.7. When I’m writing this, I can see that not only the ocfs2 disks
>> are different. I see that also the OS disks on ext3 file system is
>> different. When I check them against the nodes now, it’s all equal and
>> no differences. Has anybody experienced something like this? There was
>> no reboot between the deinstallations and installations of ocfs2.
>>
>> Mvh
>>
>> Morten Kristiansen
>>
>>
>> _______________________________________________
>> Ocfs2-users mailing list
>> Ocfs2-users at oss.oracle.com
>> http://oss.oracle.com/mailman/listinfo/ocfs2-users
>
>
>
> _______________________________________________
> Ocfs2-users mailing list
> Ocfs2-users at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-users