Hi Guys,
I'm hopping to get some help with vRA and seeding VMs. The issue I am having is when vRA starts the replication it create a new folder on the datastore with the same VM name but with a (1) at the end. The new folder it contains just replica data so the hbrxx.xxx files. If we do a failover this new (1) folder gets the .vmx, vswp, log, ect.
When I created the replication and pointed the VM at the DS it detected that there was seed files there.
VC is v6.5 and vRA is 8.1.
So;
[datastore]VmName - has the all the vmdks
new folder
[datastore]VmName(1) - gets the replica files and if failed over all the .vmx and config files.
Is there a way to seed the replication with vmdks or entire VM folders that does not result in this split
Ended up login a case with VMware. This apparently is a known bug, no fix atm.
The chances of creating a folder with (1) is , if there is another folder on Datastore with same name. Can you confirm if the datastore of source and replicated VM are different ? also check if the datastore already has another folder with same name.
There is a folder with the same name in the datastore, thats the seed vmdks. The source and destination datastores are in different DC's.
The setup is two VCs, two vRAs.
eg.
Source:
Vm Name: MyHappyVM
Datastore: LUN01
Folder on Lun: [LUN01] MyHappyVM
VMX File:[LUN01] MyHappyVM/MyHappyVM.vmx
Destination
Datastore: DR-LUN01
Seed VM Folder: [DR-LUN01] MyHappyVM
Seed VMDK:[DR-LUN01] MyHappyVM/MyHappyVM.vmdk and [DR-LUN01] MyHappyVM/MyHappyVM-ctk.vmdk ect.
Adding to this, to clarify.
Once we do a failover, the Destination config looks like this.
VMX Folder: [DR-LUN01] MyHappyVM(1)/MyHappyVM.vmx
Logs Folder:[DR-LUN01] MyHappyVM(1)/MyHappyVM001.log
Swap:[DR-LUN01] MyHappyVM(1)/MyHappyVM.vswp
but the disks are still in the original seed location
VMDK:[DR-LUN01] MyHappyVM/MyHappyVM.vmdk
And if I edit the running vm the settings show that the disks are in [DR-LUN01] MyHappyVM/MyHappyVM.vmdk
So how are you guys doing the seeding of the VMs?
What is the correct, supported method to seed the VMs that does not result in duplicate folders.
I just tired in my lab and it works perfectly. Possible causes of the split could be due to storage policy
1. Are you using any storage policy for datastores which splits the namespace and vmdk ?
2. when you configure a vm for replication can you see what do you see on the section target datastore , is the vmdk shown as "same as vmhome" in datastore column ? can you please post the screenshot what do you see ?
If possible, configure a test vm for replication and post all the screenshots here, I can compare what is the difference.
Not sure exactly where should i see the "same as source" bit?
This is 8.1 not not 6.1. It does not show this screen nor does it allow you to type/alter a path anymore.
It's not allowing me to edit, so.
It should be noted that it does auto detect the VMDKs correctly and on a multi disk VM it correctly detects all the disks.
The new (1) folder on holds the replica data, no vmdk data even after fail-over/recovery.
Ended up login a case with VMware. This apparently is a known bug, no fix atm.