VxVM data migration

[Kapil Jha]

Gurus,
need to explore various procedures I can use to sync data from one storage to other (storage to storage sync is not possible).

Scenario
Server1 storage A1 Site A
Server2 storage B1 Site B

I need to sync B1 with A1 and present same file system on Server1 and Server2.
I can present Storage B1 to Server1 and mirror the volumes, now what is the best way to break the mirror and present it as new dg on the server1 and then export import.

vg split and move we are getting licence issue, with taggin we can do but as per me its asking for downtime.

What is the best way to achieve this.

BR,
Kapil+

[Steven E. Protter]

Shalom Kapil,

Lots of ways to do this job.

dd can be done, even across a network.

If the data is in files on a cooked file system, you can sometimes do scp -rp , though database files won't work unless the database down.

fbackup, if you have it is another good way to transfer files between two SAN's.

rsync is another really good tool that with the proper command line parameters will let you sync only changed files.

SEP

[Kapil Jha]

Thanks for the reply Stevan,

I have considered those methods, and as I have some time I am trying to figure it out how can i do it from VxVM (there should be a simple method)

i got a couple of methods but do not have proper test environment to test so seeking advice :)

1)Present LUN to server1, mirror the volume and unpresent the LUN without doing anything and present the LUN to server2 and do
#vxdg -Cf import dg_name
It should work as servers are not seeing eath other and storage is different.
No Testing Env.

2)vxdg -o expand split sourcedg targetdg [volume|disk]
not sure if this would work on plex/disk level

3)same with
vxdg move sourcedg targetdg disk

4) tag the new disk which i explained in my first post.

BR,
Kapil+

[Emil Velez]

if we are talking about files whats wrong with rdist. It is standard to unix.

[gb karki]

Hi Kapil,

How are you dear.

There is multiple options.
1)use TDMF--- IBM Lisence S/F platform independent. ---Downtime require.

2)Use rsync --downtime require.

3)best option---this is suitable for VXVM.
use volume (plex) base mirroring.--
no downtime require.

If you want then call me, i can send you the steps.

Regards
Karki

[Kapil Jha]

Hi Karki,

I am fine sir, thanks for the reply.

Actually as I said above I have all those options and looking explore deep in VxVM, your method with Plex is not very feasible untill and unless you have licence to split the VG, or move disks from VG.

BR,
Kapil+

[Raj D.]

Kapil,

>Scenario
Server1 storage A1 Site A
Server2 storage B1 Site B


1.
- Study the current structure in Server1, Storage A1,

- Do you have raw volumes or only Filesystems in the diskgroups.
- Do you have cluster configured or stand alone nodes.
- How much data to migrate to Server2 storage B1
- Do you have highspeed wan/network connectivity between these two sites.



After evaluating option 1, you can decide the method going to be easy for the task.


2. options here:

- Create the same structure in : Server1 using storage A1. That is dgs and Filesystems.
- Go to site A:
Build a migration cabinet. With same storage capcity that server1 has been allocated. A migration cabinet can be a small portable box with lots of 1TB disks in an HP array, running redhat linux OS. If you have only filesystems data the task will be easy.
Connect the migration cabinet to the site A network, create the same directory structure of Server A. And copy the files using rsync.
Terabytes of data can be copied. This can be done before 1 days or 2 days of cutoff activity so that rsync can update the updated data easily with no further long time.

Once all the data synced to the migration cabinet. Send the migration cabinet to Site B, Hope fully another DataCenter or Building.
Connect the migration cabinet to the site B network, and transfer the data to the Server B 's filesystem using : # rsync -avz migration_cab:/dg01/FS01/* server_B:/FS01

(Remember the /dg01/FS01 in the migration cabinet is just a ext2 linux filesystem to hold data.)

Once all the data completed wait for cutover time, make sure you have wan access to server A and rsync work between server A and B.

During cutover run incremental rsync over the wan to sync up all the data from Server A.
# rsync -avzu server_A:/FS01/* server_B:/FS01/

Once done verify size . The rsync work can be done effectivly with help of script for large number of filesystems. Once the rsync incremental done , and verifed the appliations can be started on Server B, Site B.

If you have raw disks, there is a technique for that as well other than dd over ssh and wan.


Enjoy,
Raj.

[Raj D.]

Correction:

2. options here:

- Create the same structure in : Server2 using storage B1. That is creating same size dgs and volumes (Filesystems).

[tonyG_2]

o you have disk replication from storage A to storage B and snapclone tool on storage?

Not sure if this is what you are looking for but if disk replication exists you can snapclone the disk at the remote site and mount it as a new VG.
Pretty simple and fast solution if you have all the pieces available.

[avsrini]

Hi Kapil,

You can use the vxassist snapstart / snapshot commands to create the VxVM snapshots of the volumes and then
use vxdg split command to split the snapshot
volumes into different DG. Then deport and
import on Server2. But you may need flash
snap license for vxdg split command to work.

Could you confirm the VxVM version you are using?

Regards
Srini