• This topic has 1 reply, 1 voice, and was last updated October 30, 2020 by Jonathan S.

move tenant from one cluster to a new cluster

  • Jonathan S
    October 27, 2020 04:01:01 PM

    Hi,

    I’m looking for some insight on migrating a tenant from an existing vmware resource cluster to a new resource cluster with new storage.  Customer has a RaaS pool to replicate to;  We use vcloud director and the customer has vorg & vdc for recovery.  Ideally we would minimise the time without protection.  End goal is to migrate customer from DRaaS solution and DRaaS VDC to live execution in cloud on new VDC in vcloud on new backend cluster.  This would then bew protected by further DRaaS replication to another POD in one of our other locations.

    We’re a service provider and use the following setup:

    Vcloud director 10.0.0.16081830; Vsphere 6.7 update 3

    Customer over private link has zerto replication via on prem ZVM -> cloud ZCC – Cloud ZVM.

    Links are not heavily used and links between existing and new clusters / POD are multiple 10Gbit.

    We run zerto 7.5 update 4 (upgrading – one client preventing this at the moment…)

    So far I was thinking the following:

    Alter the location for replicated data per VPG from DRaaS location to new VDC Storage;  move / failover machine to new cloud execution location.  Reprotect to new DRaaS location on remote POD.

     

    However, this leaves the moved vpg unprotected until DRaaS as synced to new POD. Is there a better way to achieve the above with minimal impact to protection per VPG?

     

    Kind Regards,

     

    Jonathan

    Jonathan S
    October 30, 2020 08:46:38 AM

    Hi,

     

    After some testing this has been figured out.  Quick summary of process as follows:

     

    Repeat per VPG:

     

    1. Alter vpg journal and storage policy to new datastore (ensure backend access is as expected) for all machines in vpg.  Allow storage to move to new location.  No risk, Journal History maintained.
    2. Delete vpg from either source or target zvm – ensure disks left for preseed – customer now not protected.
    3. Recreate vpg on source zvm – ensure Zorg is correctly filled in as this is used to scope the search for preseed files; Must select advanced settings for machine / vpg and ensure datastore matches where the preseed was left / where the previous vpg was replicating to
    4. Browse for preseed disk per vm disk.
    5. Complete remaining vpg setup as required, click done.
    6. VPG recreated and undergoes delta sync.  40GB ~ 2 minutes. – Customer now re protected.

    Caveats:

    Provider DataStore Configuration:

    During this swing operation, the data store that is to be used for replication in the target location MUST be listed in the destination ZVM in the cloud datastores section – if not, we cannot browse this data store when trying to locate the PRESEED volume.

    Datastore Performance characteristics

    If we’re moving between volumes on an array the IO / Bandwidth limits on the volume will impact how quickly the delta sync completes.

You must be logged in to create new topics. Click here to login