I had many requests for an updated version of this paper, so the past couple of weeks I have been working hard. The paper was outdated as it was last updated around the vSphere 6.0 timeframe, and it was only a minor update. I looked at every single section and added in new statements and guidance around vSphere HA Restart Priority for instance. So for those running a vSphere Metro Storage Cluster / Stretched Cluster of some kind, please read the brand new vSphere Metro Storage Cluster Recommended Practices (6.5 update) white paper.
It is available on storagehub.vmware.com in PDF and for reading within your browser. Any questions and comments, please do not hesitate to leave them here.
- vSphere Metro Storage Cluster Recommended Practices online
- vSphere Metro Storage Cluster Recommended Practices PDF
Dennis van der Aalst says
Great to see these best practices updated. Recently we discussed vCenter HA. Can you elaborate about why this is not in the best practices. I also would like to see some kind of warning about the use of “even distribution” of VM’s in an unbalanced cluster. Besides these points: Great job.
Duncan Epping says
that advanced option (and recommendation to avoid it) is mentioned here:
When it comes to vCenter stretched deployments, this would require a full vCenter Server stretched deployment guide, which is what the vCenter Team is working on. I will refer to that when it is done, I don’t want to duplicate efforts 🙂 (plus I am far from an expert on that)
Does such a document exist that addresses NSX in a metro design?
There are some papers and blogs on this topic to be found here:
thank you for this update – waited a long time 😀
Do you know whether a kind of storage affinity rules like VM should be on Datastorecluster will be implimented? We’ve build some powercli scripts to prove wether a vm has datastores in the other datacenter so that crosstraffic exists, but i would be very happy if there would be a “built-in” way to ensure this.
Yes there’s a way, you can set affinity for VMDKs in SDRS, it was described by Frank here:
I probably should mention that explicitly in the doc, let me see if I can update it.
Great work Duncan!
We are reviewing our vMSC configs and we were pondering on HA Network Heartbeats.
When you have 2 different VLANs (routed, but each one homed in one site) for the Management Network , do you need to have additional isolation addresses configured, or the 2 different default gateways are the safest choice?
ps: thanks for all the invaluable content !!!
Sure that could work. That way you have 2 different “isolation addresses” as well, which is the result of different gateways.
I recently used the VMware compatibility guide to see if the LeftHand stretch cluster I have is supported with 6.5. It stated it was only supported for version 5.1. I then did a search to see what was supported under an iSCSI metro cluster and it stated no vendor is supported for vSphere 6.5.
Could you explain this or where I have gone wrong?
Ah just found the answer when reading the best practice guide provided 🙂
Paul Morrissey says
Your document states ” In a stretched cluster environment, only a single vCenter
Server instance is used”..But, a stretched cluster can also be configured with two vCenter’s using cross vCenter vMotion which is not reflected in the document. We have customers with this supported VMware configuration so wanted to ensure its reflected here as well.
[email protected] says
Not sure I understand the comment Paul, but a cluster can only exist in a single vCenter Server. Yes you could have vCenter in HA, but you are describing a different situation. You are describing I think a situation where you have multiple locations, each with it’s own vCenter Server and stretched storage, but the vSphere cluster isn’t stretched? This is indeed not described in the document as it is not considered to be a “vSphere Metro Storage Cluster”, but it is fully supported.