Hello, I have a Centos 6.2 clsuter with a CLVM partition on which I have a GFS2 file system. The problem rises when I make a snapshot from my FC NetAPP FAS2020. After I make the snapshot (it is a rw snapshot) of my LUN, I am not able to mount it from any of my cluster nodes, because the Physical Volume is seen two times one time on the standard LVM partition and the other time on the snapshot partition so te PV is foudn to be duplicate and I cannot mount it.
pvs
Found duplicate PV 5ZTDCmWHDH6M04nl58Wyyi3nYG8BOCRo: using /dev/mapper/mpathl not /dev/mapper/mpathi PV VG Fmt Attr PSize PFree /dev/mapper/mpathk NetCluster0 lvm2 a-- 1.67t 1020.00m
I need a hint on how to fix this problem. I need to access the snaphot volume to backup my data but I am unable to do it...
thank you very much
Rick
----- Original Message ----- | Hello, | I have a Centos 6.2 clsuter with a CLVM partition on which I have a | GFS2 | file system. | The problem rises when I make a snapshot from my FC NetAPP FAS2020. | After I make the snapshot (it is a rw snapshot) of my LUN, I am not | able | to mount it from any of my cluster nodes, | because the Physical Volume is seen two times one time on the | standard | LVM partition | and the other time on the snapshot partition so te PV is foudn to be | duplicate and I cannot mount it. | | > pvs | Found duplicate PV 5ZTDCmWHDH6M04nl58Wyyi3nYG8BOCRo: using | /dev/mapper/mpathl not /dev/mapper/mpathi | PV VG Fmt Attr PSize PFree | /dev/mapper/mpathk NetCluster0 lvm2 a-- 1.67t 1020.00m | | | I need a hint on how to fix this problem. I need to access the | snaphot | volume to backup my data but I am unable to do it... | | thank you very much
HINT: You'll need to generate a new UUID/rename the PV. (pvchange -u)
On 04/23/2012 03:44 AM, James A. Peltier wrote:
|> pvs | Found duplicate PV 5ZTDCmWHDH6M04nl58Wyyi3nYG8BOCRo: using | /dev/mapper/mpathl not /dev/mapper/mpathi | PV VG Fmt Attr PSize PFree | /dev/mapper/mpathk NetCluster0 lvm2 a-- 1.67t 1020.00m |
<snip>
HINT: You'll need to generate a new UUID/rename the PV. (pvchange -u)
I suggest to first backup original UUID to some file, then change it. It might be needed.
On 4/24/12 12:05 AM, Ljubomir Ljubojevic wrote:
On 04/23/2012 03:44 AM, James A. Peltier wrote:
|> pvs | Found duplicate PV 5ZTDCmWHDH6M04nl58Wyyi3nYG8BOCRo: using | /dev/mapper/mpathl not /dev/mapper/mpathi | PV VG Fmt Attr PSize PFree | /dev/mapper/mpathk NetCluster0 lvm2 a-- 1.67t 1020.00m |
<snip> > HINT: You'll need to generate a new UUID/rename the PV. (pvchange -u) > I suggest to first backup original UUID to some file, then change it. It might be needed.
the main problem is that the volumes I am talking about are clustered volumes and even umounting them give me troubles because looks liek I Cannot change uuid on a clustered LVM...
pvchange -u /dev/mapper/mpathk Found duplicate PV 46dU6F2rU9xqqOaWf8eihFwbdGp672lS: using /dev/mapper/mpathh not /dev/mapper/mpathk Found duplicate PV 5ZTDCmWHDH6M04nl58Wyyi3nYG8BOCRo: using /dev/mapper/mpathl not /dev/mapper/mpathi Unable to find /dev/mapper/mpathk in NetCluster0 0 physical volumes changed / 0 physical volumes not changed Internal error: Volume Group NetCluster0 was not unlocked Device '/dev/mapper/mpathh' has been left open. Device '/dev/mapper/mpathk' has been left open. Device '/dev/mapper/mpathh' has been left open. Device '/dev/mapper/mpathk' has been left open. Device '/dev/mapper/mpathk' has been left open. Device '/dev/mapper/mpathk' has been left open. Device '/dev/mapper/mpathh' has been left open. Device '/dev/mapper/mpathh' has been left open. Device '/dev/mapper/mpathk' has been left open. Device '/dev/mapper/mpathh' has been left open. You have a memory leak (not released memory pool): [0x1de12a0] format_instance [0x1de59e0] read_vg
hte main problem is that when I use pvchange -u it refuses to make operation to clsutered volume
Volume Groups with the clustered attribute will be inaccessible. Found duplicate PV 46dU6F2rU9xqqOaWf8eihFwbdGp672lS: using /dev/mapper/mpathh not /dev/mapper/mpathk Found duplicate PV 5ZTDCmWHDH6M04nl58Wyyi3nYG8BOCRo: using /dev/mapper/mpathl not /dev/mapper/mpathi Skipping clustered volume group NetCluster0 0 physical volumes changed / 0 physical volumes not changed
is there a way to remove the cluster bit from the volume itself ?
thanks
Rick
On 4/23/12 3:44 AM, James A. Peltier wrote:
----- Original Message ----- | Hello, | I have a Centos 6.2 clsuter with a CLVM partition on which I have a | GFS2 | file system. | The problem rises when I make a snapshot from my FC NetAPP FAS2020. | After I make the snapshot (it is a rw snapshot) of my LUN, I am not | able | to mount it from any of my cluster nodes, | because the Physical Volume is seen two times one time on the | standard | LVM partition | and the other time on the snapshot partition so te PV is foudn to be | duplicate and I cannot mount it. | |> pvs | Found duplicate PV 5ZTDCmWHDH6M04nl58Wyyi3nYG8BOCRo: using | /dev/mapper/mpathl not /dev/mapper/mpathi | PV VG Fmt Attr PSize PFree | /dev/mapper/mpathk NetCluster0 lvm2 a-- 1.67t 1020.00m | | | I need a hint on how to fix this problem. I need to access the | snaphot | volume to backup my data but I am unable to do it... | | thank you very much
HINT: You'll need to generate a new UUID/rename the PV. (pvchange -u)