I was playing in my lab this morning and figured I would record a demo of a new feature which is part of vSAN 6.6.1. The Performance Diagnostics feature is aimed to help those running benchmarks to optimize their benchmarks, or optimize their vSAN configuration to reach their expected goals. Note that it is a “cloud connected” feature, and in order to use this you need to participate in the Customer Experience Improvement Program. This however is enabled by default in the latest releases of vSphere. This means that data is send up to the VMware cloud, anonymous, then analyzed and the results are send back to the Web Client. Anyway, enough said, just watch the demo.
There was a question on VMTN this week about the use of the management IP’s in a “smaller” cluster as the isolation address for vSphere HA. The plan was to disable the default isolation address (default gateway) and then add every management IP as an isolation address. In this case 5 or 6 IP’s would be added. I had to think this through and went through the steps of what happens in the case of an isolation event:
- no traffic between slave and master or master and slaves (depending on whether the master is isolated or one of the slaves)
- if it was a slave which is potentially isolated then the slave will start a “master election process”
- if it was the master which is potentially isolated then the master will try to ping the isolation addresses
- if it was a slave and there’s no response to the election process then the slave will ping the isolation address after it has elected itself as master
- if there’s no response to any of the pings (happen in parallel) then the isolation is declared and the isolation response is triggered
Now the question is: will there be a response when the host tries to ping itself while it is isolated, as you need to add all ip-addresses to “isolation address” options for it to make sense… And that is what I tested. It will ping all isolation addresses. All but one will fail, the one that will be successful is the management IP address of the host which is isolated. (You can still ping your own IP when the NICs are disconnected even.) Leaving the VMs running as one of the isolation addresses responded.
In other words, don’t do this. The isolation address should be a reliable address outside of the ESXi host, preferably on the same network as the management.
I had this question a couple of times the past month so I figured I would write a quick post. The question that was asked is: Can I have an unbalanced stretched cluster? In other words: Can I have 6 hosts in Site A and 4 hosts in site B when using vSAN Stretched Cluster functionality? Some may refer to this as asymmetric or uneven. Either way, the number of hosts in the two locations differ.
Is this supported? In short: Yes.
The longer answer: Yes you can do this, this is fully supported but you will need to keep your selected PFTT (Primary Failures To Tolerate), SFTT (Secondary Failures To Tolerate) and FTM (Failure Tolerance Method) in to account. If PFTT=1 and SFTT=2 and FTM=Erasure Coding (RAID5/6) then the minimum number of hosts per site is 6. You could however have 10 hosts in Site A while having 6 hosts in Site B.
Some may wonder why anyone would want this, well you can imagine you are running workloads which do not need to be recovered in the case of a full site failure. If that is the case you could stick these to 1 site. (You can set PFTT=0 and SFTT=1, which would result in this scenario for that particular VM.)
That is the power of Policy Based Management and vSAN, extremely flexibility and on a per VM/VMDK basis!