cancel
Showing results for 
Search instead for 
Did you mean: 

any change on the veritas cluster after the data migration to new disks

Cluster: Veritas Cluster Server 5.0

OS: Solaris 10

Volume Manager: Symantec storage foundation 5.0

 

I have migrated the data to new disks. I would like to double check what need to be changed on the cluster end.

 

1. Shall I change the block device in the main.cf file since new disks have different device names?

    If yes, does it need to be changed after bringing down the cluster?

2. I don't know what needs to be done on the standby node.

   I guess that I don't need to anything as long as I see the shared disks.

 

Please help me.

Thank you in advance.

 

1 Solution

Accepted Solutions
Accepted Solution!

You may not have to change anything...

Block device names under Volume Manager are made up of the VxVM disk group name and the VxVM volume name.  So if I have a file system on volume vol1 in disk group mydg, it would have a block device name of /dev/vx/dsk/mydg/vol1.  So if you migrated to new disk devices but kept the same disk group and volume names, the block device names will remain the same.

View solution in original post

5 Replies
Accepted Solution!

You may not have to change anything...

Block device names under Volume Manager are made up of the VxVM disk group name and the VxVM volume name.  So if I have a file system on volume vol1 in disk group mydg, it would have a block device name of /dev/vx/dsk/mydg/vol1.  So if you migrated to new disk devices but kept the same disk group and volume names, the block device names will remain the same.

View solution in original post

HI, COuld you please let me

HI,

COuld you please let me know what process/steps you used to migrate the data to new volumes.   I have the same work scheduled the same.

Regards,
GOvinda.

 mirror the volume first,

 mirror the volume first, once they are fully synchronized, split the mirror and delete the first plex.

Through this approach, you don't need to do anything to the cluster.   every singe operation is manually conducted outside of VCS.

Really?

Eric,

So do you just have to mirror the volumes, split the mirrors and delete the first plex?  I'm familiar with doing this for a single Sun server, but not for a cluster.  I haven't done it for a cluster before.

Does this still hold true if I have an active/active cluster?  Does VCS just propogate the changes out to the other nodes?

Thanks.

Hello, You need not to change

Hello,

You need not to change anything at all.... let me show you with an example, possibly will help you understanding it....

Lets say you have a volume configured in cluster like below:

Mount Mount_oraprod_u01 (
        MountPoint = "/prod/u01"
        BlockDevice = "/dev/vx/dsk/ora_prod_dg/u01-vol"
        FSType = vxfs
        FsckOpt = "-n"
        )


& above volume looks like this in volume manager

v  u01-vol     -            ENABLED  ACTIVE   62914560 SELECT    -        fsgen  <<<<<<<<<<<<< this is your volume
pl u01-vol-01  u01-vol     ENABLED  ACTIVE   62914560 CONCAT    -        RW
sd u01-voldg-01 u01-vol-01 u01-voldg 0 62914560 0        EMC0_31  ENA

Now once you mirror it under vxvm, it should be like this:

v  u01-vol     -            ENABLED  ACTIVE   62914560 SELECT    -        fsgen  <<<<<<<<<<<<< this is your volume
pl u01-vol-01  u01-vol     ENABLED  ACTIVE   62914560 CONCAT    -        RW
sd u01-voldg-01 u01-vol-01 u01-voldg 0 62914560 0        EMC0_31  ENA
pl u01-vol-02  u01-vol     ENABLED  ACTIVE   62914560 CONCAT    -        RW
sd u01-voldg-02 u01-vol-02 u02-voldg 0 62914560 0        EMC0_32  ENA

Now you can see above, even after mirroring, your volume name still remains same, nothing has changed to volume...

Now if you remove your first plex....

v  u01-vol     -            ENABLED  ACTIVE   62914560 SELECT    -        fsgen  <<<<<<<<<<<<< this is your volume
pl u01-vol-02  u01-vol     ENABLED  ACTIVE   62914560 CONCAT    -        RW
sd u01-voldg-02 u01-vol-02 u02-voldg 0 62914560 0        EMC0_32  ENA


Still your volume remains same & block device path is still same i.e /dev/vx/dsk/ora_prod_dg/u01-vol

So no change is needed in cluster config. Also, cluster config is same on active  & passive node, so no different for passive node as well..

Hope this helps..


Gaurav