2V0-620 · Question #146
An administrator is attempting to migrate a large Hadoop virtual machine using vMotion. The administrator notices a delay when the machine is quiesced. The Hadoop virtual machine is processing many…
The correct answer is A. Use multiple NICs for the vMotion vmkernel port. B. Use a 10Gbps Network card for the vMotion vmkernel port. The vMotion quiesce phase is delayed when a VM has a high memory dirty rate; increasing network bandwidth through multiple NICs or a faster NIC reduces the time needed to transfer and synchronize memory pages.
Question
An administrator is attempting to migrate a large Hadoop virtual machine using vMotion. The administrator notices a delay when the machine is quiesced. The Hadoop virtual machine is processing many transactions a second to an in-memory database. Which two actions would help reduce the amount of time needed to perform a vMotion on this virtual machine? (Choose two.)
Options
- AUse multiple NICs for the vMotion vmkernel port.
- BUse a 10Gbps Network card for the vMotion vmkernel port.
- CAdd an additional management network to help transmit the data quicker.
- DDisable Fault Tolerance before performing the vMotion.
How the community answered
(66 responses)- A70% (46)
- C21% (14)
- D9% (6)
Why each option
The vMotion quiesce phase is delayed when a VM has a high memory dirty rate; increasing network bandwidth through multiple NICs or a faster NIC reduces the time needed to transfer and synchronize memory pages.
Using multiple NICs for the vMotion vmkernel port enables NIC teaming and aggregates available bandwidth, allowing the large volume of in-flight memory pages from the high-transaction Hadoop VM to be transferred faster and shortening the quiesce (stun) phase.
A 10Gbps NIC provides up to 10x the throughput of a 1Gbps NIC, which directly reduces the time required to transfer memory state during vMotion and minimizes the duration of the quiesce phase for a VM with a high memory write rate.
The management network handles vCenter-to-host control traffic and has no role in transferring vMotion memory data, so adding a management network does not improve vMotion throughput.
Fault Tolerance prevents vMotion entirely rather than causing a delay - if FT were enabled, the migration attempt described in the question would not be possible at all, not merely slow.
Concept tested: vMotion network bandwidth optimization for high memory workloads
Source: https://docs.vmware.com/en/VMware-vSphere/6.7/com.vmware.vsphere.vcenterhost.doc/GUID-A27EE017-F186-4DB5-8804-E5E0CEDE8D88.html
Topics
Community Discussion
No community discussion yet for this question.