In the introduction to this series, we identified that there are multiple parts of the cloud migration journey; namely: Discover, Assess, Migrate, Manage, and Operate.
We will cover various tools that either falls into one of or span multiple phases in the migration journey.
As a quick-link reference, here is the list of tools this series will cover:
- Part 1: Introduction
- Discover
- Part 2: Microsoft Assessment and Planning (MAP) Toolkit
- Part 3: Service Map
- Part 4: OVF/OVA
- Part 5: Custom Scripts
- Assess
- Migrate
- Part 10: Azure Websites Migration Assistant
- Part 11: Azure Database Migration Planner
- Part 12: Azure Migrate
- Part 13: Azure Migration Center
- Part 14: Manage and Optimize
- Part 15: Third-Party Tools
- Part 16: Summary
Continuing our focus on the tool(s) that fall within the second phase, Assess, the next tool that we categorize under Assessment, is the Azure Site Recovery (ASR) Deployment Planner Tool.
This is a command-line tool that has a 4-mode process to its use.
First, you need to either obtain or provide a list of all the Virtual Machines you want to profile.
Second, once a list of target Virtual Machines is obtained, the tool will need to initiate profiling against the VMs. This consists of connecting to each Hyper-V host (or vCenter/ESXi host) and collecting performance data. It will query the host once every 15 seconds and stores the average of every minute’s performance counter data collected.
Third is Virtual Machine profiling. The challenge with this tool is the profiling duration. The minimal profile duration is 30 minutes; however, the recommendation is to run profiling for at least 7 days, with a “best” official suggestion of 31 days! This, of course, is to ensure that the workload pattern in the environment over the specified period is observed so that it can be used to provide an accurate recommendation.
To measure replication throughput, you must create an Azure Storage account, so that measurements can be taken between on-premises and Azure Storage. It does this by creating multiple temporary VHD files (each being 64 MB in size) and uploads them into the target Azure Storage account to determine the achievable throughput. After the throughput is calculated, it will delete the temporary files from the both the Azure Storage account, and the local copies.
The fourth step is to generate the report based on the data collected.
What Is It Used For?
This tool is used for profiling your Hyper-V and VMware virtual machines, to provide supported Disaster Recovery recommendations, factoring in network bandwidth, storage, replication, data churn, IOPS, CPU/RAM, etc.
The report that is produced can provide details including:
- Required network bandwidth to achieve target RPO
- Number of Azure Storage accounts required
- Number of Azure Cores required
- Which Virtual Machines are compatible/incompatible
- Total DR costs per month
- Azure Storage costs per month
Pro’s vs Con’s
Here is a quick list of what we deem as Pro’s/Con’s of this tool.
Pro’s
- The same tool supports both VMware and Hyper-V to Azure disaster recovery scenarios
- If targeting a Hyper-V cluster, the tool finds all the Hyper-V nodes in the cluster and obtains the VMs present on each of the Hyper-V hosts
- Able to adapt to VM migration from one node to another in a cluster
- If the server that is running the profiling tool is rebooted, or the tool has crashed, or the tool has been closed, profiled data is preserved
- Generated report can include growth factor (default is 30%)
- The percentile value of 95 for read/write IOPS, write IOPS, and data churn is used, in order to ensure that temporary spikes are not used to determine target storage account and source bandwidth requirements
Con’s
- The tool must be run on a Windows Server 2012 R2 or Windows Server 2016 system
- ASR Deployment Planner is a command-line tool only
- You must run 4 separate command modes independently (i.e. there is no single command that will initiate all modes as needed sequentially)
- Recommended profile duration is 30 days (minimum is 30 minutes)
- Throughput is calculated from the server where the tool is running to Azure, not from the target profiled VMs
- Any VM configuration changes (i.e. CPU, Disks, NICs) applied after the profiling has been initiated are not captured
- VMs that have iSCSI, pass-through disks, and NFS disks are not supported (and profiling does not capture data on these disk types)
- VMware vCenter statistics level settings need to be set to 2 or high level
Other Thoughts
Although the tool takes a long time to collect data, it is very detailed in what it collects. With such a large volume of data, sifting through all the data can be overwhelming.
The Recommendations report that is generated simplifies this and provides nice visuals and easy to understand data points and even gives you some estimated Disaster Recovery costs.
It is important to note, that even though the Azure Site Recovery (ASR) Deployment Planner Tool supports both Hyper-V and VMware, it is highly recommended to use the new Azure Migrate tool (which will be covered in another part of this series) for VMware environments.