Exam 70-740 topic 1 question 116 discussion

Half here say C the other D, which is correct?

With ResiliencyPeriod, you decide how long this node stays in isolation before it's deemed a lost cause, and the cluster fails the node. A setting of 0, that is, 0 seconds, means that a node fails immediately REGARDLESS of the cause. It does not matter what the setting is for ResiliencyLevel, because both settings (Level 1 and 2) would Isolate the node for a Cluster Service crash, so setting the Level itself is useless. Thus the answer is 10000% "C. ResiliencyPeriod."

With ResiliencyLevel, you can select 1 or 2. For Level 1, the node will Isolate (chill-out) if the Cluster knows that whatever happened to it is not severe, so to speak, aka a known reason. Otherwise it goes back to a trusty ol' failure. Level 2 always isolates first, and waits for a time before actually failing it, regardless of the cause. This is the default setting. Here comes the second setting: ResiliencyPeriod. --Cont.

Question seems wrongly worded. Regardless: Isolation is Different from Failure. They are not the same thing. The entire premise of these settings is that new Server 2016 concept: Isolation (And Quarantine, but that is not relevant here.). This state allows a VM to chill the f out instead of starting a potentially more-troublesome failover. It's not possible to remove this new setting, but you can modify it. The Question was very specific about the error we are dealing with, that is, a Cluster Service crash.--Cont.

Correct answer is D ResiliencyLevel - Defines how unknown failures handled 1 – Allow the node to be in Isolate state ONLY IF THE NODE GAVE NOTIFICATION and it went away for known reason, OTHERWISE FAIL IMMEDIATELY. Known reasons include Cluster Service crash or Asymmetric Connectivity between nodes.

I should also say that the resiliency level is irrelevant as either setting 1 or 2 will work if the cluster service goes down. So you just need to set the resiliency period.

The answer is C. The default resiliency level is 2 - which means that a node will always go into an isolated state. A resiliency level of 1 would mean that for known issues - such as the cluster service crashing - it would allow the node to stay in an isolated state. So you do not need to change the level. However you do need to change the resiliency period to 0 so that it fails over immediately. The default is 240 seconds.

I think C and D are correct. However, the question needs to say as follow. "Which two settings should you configure ?" or "Which setting should you configure first ?"

Answer is C: https://techcommunity.microsoft.com/t5/failover-clustering/virtual-machine-compute-resiliency-in-windows-server-2016/ba-p/372027 ResiliencyPeriod 0 – Reverts to pre-Windows Server 2016 behaviour If you set it to 0 the VM will failover immediately.

I think C is correct. --D.ResiliencyLevel It specifies the kind of fails, such as known or unknown fail. The requirments for this question is to fail over immediately, but it can't specify that action. --A.FailureConditionLevel Since the question says that VMs are configured to fail over...., it has been set already.

I believe the answer should be A. Configuring the conditions for failover with the failover policy. https://docs.microsoft.com/en-us/sql/sql-server/failover-clusters/windows/failover-policy-for-failover-cluster-instances?view=sql-server-ver15 Anyone can confirm?

seems right: https://techcommunity.microsoft.com/t5/Failover-Clustering/Virtual-Machine-Compute-Resiliency-in-Windows-Server-2016/ba-p/372027 ResiliencyLevel - Defines how unknown failures handled 1 – Allow the node to be in Isolated state only if the node gave a notification and it went away for known reason, otherwise fail immediately. Known reasons include Cluster Service crash or Asymmetric Connectivity between nodes. 2- Always let a node go to an Isolated state and give it time before taking over ownership of the VMs.

I tend to agree https://techcommunity.microsoft.com/t5/Failover-Clustering/Virtual-Machine-Compute-Resiliency-in-Windows-Server-2016/ba-p/372027