Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations Mike Lewis on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

expand existing vxvm based fs (with vcs, vxvm,emc) 1

Status
Not open for further replies.

babcia01

IS-IT--Management
Jul 11, 2002
37
US
Hi,
Could you please point me to some examples or good reference manuals to perform the following:

- add new LUN EMC to existing dg and expand size of existing vxvm based file system which is part of the veritas cluster by x GB.

To summarize:
The environment consist of:
Solaris OS
Veritas Cluster
VXVM LVM and
EMC disk subsystems running POWERPATH.

New LUN EMC disk will be presented and needs to be added to existing disk group. The lvm is vxvm and dg is part of the clustered environment.

Can such expansion "normally" or most often be performed dynamically without the need to bring the veritas cluster down?

I did the search on the Web and did not find good source yet.

Thank you in advance
 
Once the disk is visible at the OS level, everything can be done on-the-fly with no impact to service or VCS. The awkward bit can be making the disk visible at the OS level, especially under Solaris with it's inflexible sd driver, unless you have preempted this situation by creating free sd 'slots'. What version of Solaris?

Once the disk is presented, and presuming the preparation above has been done, the procedure should be roughly:

1. Present the disk to the host across the SAN.

2. Scan for the new disk and create /dev entries:

[tt]devfsadm -c disk[/tt]

3. Add the disk to PowerPath:

[tt]powermt config[/tt]

3. Label the disk using format.

4. Initialize the disk for use by VxVM:

[tt]vxdisk init cXtXdX [ privlen=private_region_length ][/tt]

or on older versions:

[tt]vxdisksetup -i cXtxdX [ privlen=private_region_length ][/tt]

5. Add the disk to the disk group:

[tt]vxdg -g diskgroup adddisk diskname=cXtXdX[/tt]

6. Increase the volume size:

[tt]vxassist -g diskgroup growto volname newsize diskname[/tt]

7. Increase the filesystem size (note that the size here has to be in 512-byte blocks, can be obtained from vxprint output):

[tt]fsadm -F vxfs -b newsize /filesystem/mountpoint[/tt]

6 & 7. Or, alternatively, if you prefer to combine these steps:

[tt]vxresize -bxF vxfs -g diskgroup volname newsize[/tt]

Annihilannic.
 
Hi Annihilannic
Thank you for the reply. The OS is 5.9.
Did your reply take into consideration that the dg is part of the veritas cluster? Anything that should be done to account for the change to be reflected on the other cluster node so that OS knowledge on each of the cluster node is applied and consistent to allow for a successful package failover?
 
If the disk group is a shared disk group, i.e. can be imported simultaneously on both nodes of the cluster, the change will be reflected immediately on both nodes. Note that in this scenario you will need to perform the VxVM steps on the Cluster Volume Manager master node (use vxdctl -c mode to check which one that is).

If it's a standard 'failover' type disk group, then the changes you make will be stored in the VxVM configuration for that disk group (which is in the private region on the disks belonging to that group), so it will travel with the disk group to whatever node you import it on.

Annihilannic.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top