delete remote copy group with boot disk

[RiclyLeRoy]

I wish to delete remote copy group which collects 4 virual volumes and they are boot disks of 4 blade servers.

Remote copy group replicates data from 3par 8200 (A: called source storage) to another one 8200 (B: called destination storage)

My target it's to delete the remote copy group, keeping storage B and to not use more storage A, to change configuration of each server to point directly to B virtual volume. What steps do you suggest to follow?

[sbhat09]

Hello @RiclyLeRoy,

This is the high level steps for the configuration changes you are looking for:

1. Stop the replication
2. Ensure that exported volumes from the site B are fully functional as boot volumes for the hosts
3. Ensure there is no data descrepancy between source and destination (If you have any tests to run, please run and take confirmation)
4. From the source (Site A) delete the remote copy group
5. The volumes in site A and site B will work as independent volumes

Note: Deleting the remote copy group will only delete the logical grouping of the volumes. It will not impact individual volumes or the data inside the volume.

You find this discussion chain helpful as well - https://community.hpe.com/t5/hpe-3par-storeserv-storage/hpe-3par-simulator/m-p/7099892#M6031

Let me know if you need more details on this.

Regards,
Srinivas Bhat

If you feel this was helpful please click the KUDOS! thumb below!
Note: All of my comments are my own and are not any official representation of HPE.

[Cali]

I don't think it will work the way Srinivas writes. (Sorry Srinivas)
First, you must clarify whether it is a sync or async replication.
When you delete the rcgroup, both volumes are retained, but both are also actively exported to the server, leading to problems.
So with sync, I would first do a switchover to B, so that the active paths come now from B.
Then you can unexport the passive paths from A and only then delete the rcgroup.
And if you have async replication, it is more complex.

Cali

[RiclyLeRoy]

remote group replication is sync.

I thought these steps:

poweroff the server

stop replication

on server, to change the pointing of boot HBA device to B virtual LUN

remove vlun from replication

unexport vlun on 3par A (source)

destroy replication

poweron server

what do you think ?

[Cali · Accepted solution]

>> on server, to change the pointing of boot HBA device to B virtual LUN

The primary and secondary LUN have the same Volume Identifier and LUN ID, otherwise, a switchover will not work.

So, you can not point to B-LUN, because it has the same ID.

>> remove vlun from replication

>> unexport vlun on 3par A (source)

First unexport from A and then remove from replication.

Cali

[RiclyLeRoy]

>> First unexport from A and then remove from replication.

Why I have to unexport vv from A ? In this way all changes from host to source volume will be stopped ? I thought the same thing is get when host is power-off.

My host has n.2 FC ports and in boot menu of server there 2 entries:

(WWN...2..,Lun0000)/3PARdataVV

(WWN...1..,Lun0000)/3PARdataVV

What WWN are ? vLun ?

>>The primary and secondary LUN have the same Volume Identifier and LUN ID, otherwise, a switchover will not work.

How can Verify by 3par console that volume identifier and LUN id of A and B LUN are the same value ?

Thanks

[RiclyLeRoy]

I unexport A virtual LUN but I try to remove this vLUN from copy group I get this errror:

Error: command failed on target system...,

returned error: Volume xxx of group yyy is a Peer Persistence volume. Dismissal of this volume from the group will cause the LUN Alua state to go 'Active'. Please unexport the secondary Peer Persistence volume or disable path_management policy for the group and retry.

[Cali]

Switchover first and than Unexport.

[RiclyLeRoy]

what I have to switchover ?

[Cali]

A has the Active Path

B has the Passive Path

If you Unexport A, you loose the Active Path.

Look in

SSMC > Volumes > Select the Volume  > Export

and see also

SSMC > Volumes > Select the Volume.r  > Export

If you Switchover this Volume, it is Active on B and Passive on A.

Cali

[RiclyLeRoy]

Copy group has several volumes.

I think I can switchover the whole copy group, I am not able to choose a single volume inside copy group.

right ?

[Cali]

OK, in that case, I would turn off the server with access to the volume, then unexport from A.
Remove the volume from the rcgroup (ignore the error).
After that, the paths on B should go from standby to active and you can start again.