vRep 6.1.1 - replication questions for modifying virtual machines in vRealize workflows

[sparkblaze]

In the web GUI, it's possible to see the MPiT data for replicated VMs (instances per day etc), however, this doesn't seem to be exposed at all to vRealize Orchestrator, PowerCLI or vim-cmd hbrsvc.

Is this at all possible to obtain from any of these methods, or indeed, any method via scripting?

Alongside this, when you do things like expand a VMDK attached to a replicated VM, as you have to move the folder and stop replication, when you restore the folder and re-create the replication (manually in this instance), using the original as the seed.... it doesn't re-attach the already created point in time instances - can they be re-attached? should they be? if not, will it tidy itself up?

[mvalkanov · Accepted solution]

Hi,

Information on the MPITs is available only at the target site VR server (hbrsrv) and thus is not available through source ESXi vim-cmd hbrsvc.

The VRMS API is not public. It is possible to browse available PITs using that API but through Managed Object Browser - still this would be manual.

I'm not sure if the vRO plugin for VRMS exposes this or not, but should be easy to implement - perhaps you can file a feature request for that.

About resizing replicated VMDK - the VR server uses a chain of redolog-based snapshots at the target site - a base disk and each delta stored as a separate child disk. This format of "redolog-based snapshots" does not allow different sizes of the different PITs. This is planned to be addressed by a new vSphere snapshot format (quickly looking at Internet this seems to be not yet publicly announced). So short answer is no - the old PITs would be lost. Unless you do a recovery without power on and clone the recovered VM, before using its disks as initial seed for the new replication (w/ resized disk).

Regards,

Martin

[sparkblaze]

hmmm, I was afraid that this would be the answer.

How would I use the VRMS API via the MOB? I had a look at the target vcsa and replication host, but couldn't find anything relating to vrms/hbr or replication. I'll look at submitting a feature request regardless, but I imagine VMware would rather people upgrade to SRM.

I'll keep an eye out for the new snapshots then, but I guess that be a feature for v7 if not later.

[mvalkanov]

Hi,

To list all incoming replications:

https://vrms_at_target_site:8043/mob/?moid=replica-manager&method=HmsReplicaManagerGetIncomingReplic...

start=0

count=2000

sorters - clear everything from the textbox

filters - clear everything from the textbox

Note the GID_... values - these are the IDs of the incoming replications.

For such a value - to list all available PITS:

https://vrms_at_target_site:8043/mob/?moid=replica-manager&method=HmsReplicaManagerFindGroupInstance...

<replicationGroup type="HmsGroup">GID-e91565d5-40a0-495c-b1e5-70ca53f08a78</replicationGroup>

SRM does not provide any more support on multiple PITs with VR - only be able to recovery to one of those PITs.

Better MPIT story for VR is on the roadmap for a future release, but I can not share specific timeline.

Regards,

Martin

[sparkblaze]

Odd, these links work, but when I look in replica-manager it lists no methods or properties.

I have since discovered the text file in the target replication datastore that also gives me the info I want though.

hbrgrp.GID-[GUID].txt contains the group-wide state info and the instance frequency over days info can be worked out from the group.rpo and group.retentionPolicy values.

I've dished out the relevant best answer as well.