Shulou(Shulou.com)06/03 Report--

The following mainly brings you the cross-cluster migration of WSFC from 2008R2 to 2012R2. I hope these contents can be of practical use, which is the main purpose of my article on the cross-cluster migration of WSFC from 2008R2 to 2012R2. All right, don't talk too much nonsense, let's just read the following.

The environment is as follows

2008R2 cluster

Node1

MGMET:10.0.0.3 255.0.0.0 DNS 10.0.0.2

ISCSI:30.0.0.3 255.0.0.0

CLUS:18.0.0.3 255.0.0.0

Node2

MGMET:10.0.0.4 255.0.0.0 DNS 10.0.0.2

ISCSI:30.0.0.4 255.0.0.0

CLUS:18.0.0.4 255.0.0.0

Cluster Information Name:devtest.oa.com IP:10.0.0.99

08dc&iscsi

Lan:10.0.0.2 255.0.0.0

Iscsi:30.0.0.2 255.0.0.0

12R2 cluster

HV01

MGMET:10.0.0.9 255.0.0.0 DNS 10.0.0.2

ISCSI:30.0.0.9 255.0.0.0

CLUS:18.0.0.9 255.0.0.0

HV02

MGMET:10.0.0.10 255.0.0.0 DNS 10.0.0.2

ISCSI:30.0.0.10 255.0.0.0

CLUS:18.0.0.10 255.0.0.0

Cluster Information: name:pecluster.oa.com ip:100.0.0.100

As you can see, the current DHCP cluster role works on top of the 2008R2 cluster, named devtestdhcp

Dhcp is configured with scope and specific retention options

Cluster storage is allocated as follows: 08R2 has a separate set of cluster storage, and 12R2 has a separate set of cluster storage. In this example of DHCP migration, we directly introduce the scenario of cross-cluster migration and different cluster storage.

The migration time node is cut as follows

The DHCP cluster group runs on the 2008R2 cluster to provide services

At this time, the client DHCP obtains the address and responds from the 2008R2 cluster.

Create a 2012R2 cluster, and a 2012R2 cluster connects to new storage

Run the Cross-Cluster Migration Wizard to migrate the DHCP cluster group to the 2012R cluster, and the cluster group is offline after the migration

Back up the source 2008R2 cluster DHCP configuration

Offline source 2008R2DHCP cluster group

Online 2012R2DHCP cluster group

Restore DHCP configuration

The client obtains the DHCP address from the 2012R2 cluster

The first three steps are omitted. Starting from step 4, run the cross-cluster migration wizard. In 2012R2, the cross-cluster migration wizard has been renamed to replicate the cluster role, but the function remains the same and is still familiar.

It is important to note that 2012R2's Cross-Cluster Migration Wizard only supports the migration of roles from the cluster of 2008R2Magol 2012R2 if your cluster environment is 2008Magi 2003. Please migrate to the 2008R2 cluster first. at the beginning of 2008R2, when we call a cluster group, it is a friendly way to say the cluster role directly, and to the technicians, we know that the cluster group is the cluster role, and the cluster group is the smallest failover unit in the cluster.

Enter the 2008R2 cluster name

Select the appropriate cluster role. Here we will only select the dhcp cluster role.

Disk mapping, mapping the cluster disk of the DHCP cluster group in 2008R2 dhcprole configuration to the cluster disk dhcp2012 in 2012R2

Click next, you can see that it is pointed out in the confirmation summary report that the wizard will copy the DHCP cluster role configuration and disk configuration to the 2012 R2 cluster. From 2012, the concept of the replication cluster role has changed. The original 2008R2 is called migration service and application, but the source side is still there after you actually migrated, so 2012 is called replication from the beginning, allowing the cluster role and disk configuration in the old cluster to be copied to the new cluster.

Then run into the degree bar, wait, this process is usually very fast, because this is a light wizard, it only copy configuration, not the actual content migration, after completion, we will get a summary report, if you want to see the details, you can also check the detailed report, will point out the results of the various content migration, as well as the follow-up work of the migration.

At this point, you can see that the DHCP cluster role and configuration have been replicated to the 2012R2 cluster and are currently offline

Since we need to switch off the old cluster DHCP cluster role and launch the new cluster DHCP role, we need to perform a backup of the old cluster DHCP configuration now.

# execute the export command, we need to use the netsh command to export before the migration from 03 to 08:00, but there has been a better PS command since 2012, which we use directly on the 2012R2 node to connect to the DHCP2008R2 cluster role export

Export-DHCPServer-ComputerName devtestdhcp-Leases-File C:\ 08dhcp.xml-Verbose

Offline 2008R2DHCP cluster role resources, at this time, if the client tries to obtain the IP address, it will not get it, so this part is downtime and requires the administrator to plan the time node

Online 2012R2DHCP cluster role

At this time, although the cluster role has been online, but open the DHCP manager of 2012R2, you can see that there is no reserved option, the previous configuration has been lost, and is currently operating as a blank database

We need to import the previously exported 2008R2DHCP cluster configuration into 2012R2

# Import command backpath can specify an existing path

Import-DHCPServer-ComputerName devtestdhcp-File C:\ 08dhcp.xml-Verbose-BackupPath C:\ dhcpbackup

When you open the DHCP manager again, you can see that the configuration data has come. A key point here is the authorization of DHCP. When migrating from stand-alone to stand-alone DHCP, we need to plan DHCP authorization reasonably. Only authorized DHCP Server can provide DHCP services. So when we migrate from stand-alone to stand-alone, we will revoke the authorization of the original computer and authorize the new computer. And DHCP migration will not involve this problem, because devtestdhcp this cluster computer to provide DHCP services, when we operate in 2008R2, we will first authorize the computer in AD before we can provide DHCP services, so when the whole object is provided by 2012R2, it has already been authorized, so there is no need for re-authorization, the role online can provide services!

At this time, the client obtains the IP address again and finds that it can be obtained successfully. The DHCP CVM is still the same address and the configuration is the same, but it is already provided by the 2012R2DHCP cluster behind the scenes!

Manually moving the DHCP cluster role to the Node2,DHCP role still works

At this point, we have completed the smooth migration of DHCP clusters from 2008R2 to 2012R2. Lao Wang will no longer demonstrate the method of using old storage for both new and old clusters, because they are all the same, exchanging cluster roles and configurations, but missing the step of disk mapping. In the end, we still have to back up and restore the DHCP configuration manually.

Some friends may say, is it necessary? why does DHCP still want to do clustering? in fact, there are still many ways for the network service DHCP to be highly available now. Split the scope, 50 pounds, 50 pounds, 80 pounds, 20 percent, and 2012DHCP natively supports failover. DHCP clusters are actually used more often in the 03 and 08 times, because some enterprises also use Windows Server as a DHCP cloud server. Once used, of course, I want a highly available service at this time, so I will choose to deploy a cluster, so a set of cluster databases will be deployed. One of my nodes is down, and the other can continue to provide DHCP services. If the old DHCP cluster is configured with many options, then cross-cluster migration makes sense.

In the traditional cluster role, Lao Wang thinks that DHCP and file cloud server are two typical ones, so he chooses to talk about it. This article will also mention virtual machines. In this experiment, Lao Wang deployed both DHCP and virtual machines on the nodes, which is not recommended in practice. In practice, it is recommended that virtualized cluster nodes only provide virtualization and do not assume other cluster roles. Lao Wang hopes to tell you more here, so he deployed both functions.

OK, next, the most important part of the migration series, whether it is possible to migrate virtual machines across clusters, the answer is yes.

First of all, let's take a look at the centralized general solution for virtualization migration upgrade.

If you want to upgrade 2008R2's Hyper-V virtual machine to 2012R2, there are roughly the following ways

No cluster

1. Shut down and export the virtual machine and import it to 2012R2

two。 Shutdown, manually copy virtual machine configuration and virtual machine disk, import to 2012R2

Therefore, you can see that in the case of no cluster, the upgrade of the 2008R2 virtual machine to 2012R2 is bound to be down.

After export, import on the 2012R2, and then determine that there are no virtual machines on the original Hyper-V node, clean reinstall, upgrade to 2012R2, and then migrate the virtual machines back in real time. It seems that there is only this route to choose when there is no cluster. If you do it in a production environment, you must choose the maintenance time.

If you adopt the import and export scheme to upgrade your virtualization, there are a few points to pay attention to

Ensure that the exported virtual machine has no save state and no snapshot

Make sure that the export virtual machine MAC address is set to static

Before 2012, the 2008R2 virtual machine must be in the virtual machine shutdown state to be exported.

It is best to back up the exported virtual machine

Try to use standard network adapters for export virtual machines instead of legacy network adapters

After the load import and export on the node is migrated away, it is recommended to install it cleanly instead of upgrading the installation.

There are clusters

No VMM.

Ensure that both new and old clusters can connect to storage, but storage is only visible to a single cluster at the same time

Storage is not added to the new cluster disk

Run the Cluster Migration Wizard, migrate the virtual machine configuration, CSV configuration to the new cluster

Offline old cluster virtual machines, unassigned disks to the old cluster

The CSV migrated from the old online cluster automatically uses the disks previously assigned to both the old and new clusters

Online virtual machine

With VMM, you can manage 2008R2 clusters and 2012R2 clusters at the same time through VMM. You can directly operate cross-cluster migration in the VMM console, and directly move virtual machines from 2008R2 clusters to 2012R2 clusters.

VMM supports upgrading from 2008R2 cluster to 2012R2 cluster, 2008R2 cluster to 2012R2 stand-alone, and 2008R2 stand-alone to 2012R2 cluster, but it takes a little longer to upgrade using VMM than using the cluster migration wizard, because the actual VMM performs an export import operation for the old version to upgrade to the new version.

From this, you can see a problem. If we have a cluster, the downtime of our migration process is very short. The virtual machine does not need to be shut down all the time to perform the migration. We can keep running the virtual machine until we have to exchange resources with 2012R2, go offline, and there will be a little downtime when we go online, but this time is much shorter than the downtime of shutdown, export and import.

In the migration of virtualization, there is also a scenario, that is, storage mapping. If we want to perform virtual machine migration across clusters, we must CSV the virtual machine to store that disk and map it to both 08R2 and 12R2 clusters, but cut the time node, 12R2 does not add storage to the cluster until 08 is offline, 12R2 goes online because the virtual machine is migrated, it will carry the CSV attribute, and CSV only recognizes that disk. If we adopt the new cluster and the new storage, the old cluster will lose the CSV and it will not be able to recognize the disk it is using.

When we use the cross-cluster migration wizard, we need to note that for the cross-cluster migration wizard for virtual machines, the migration wizard essentially migrates according to CSV. We can only select one CSV, and then migrate all the virtual machines below, instead of selecting a specific single virtual machine under the CSV. Therefore, we can see that the cross-cluster migration wizard is suitable for scenarios with many virtual machines running above the cluster. If there are not many virtual machines, you can also choose to import and export, but Lao Wang still recommends to use the cross-cluster migration wizard because this has the shortest downtime.

Let's verify it by experiment.

Virtualization across cluster migration nodes is cut as follows

The virtual machine operates on the 2008R2 cluster CSV

Assign free disk targets behind CSV to 2012R2 cluster nodes

The 2012R2 cluster node can see the disk, but does not add to the cluster storage

Perform the Cross-Cluster Migration Wizard to migrate virtual machines

Offline 08R2 virtual machines, storage unallocated to 08R2 cluster nodes

12R2 Cluster online CSV

12R2 clustered online virtual machine

Allocating 08 storage is visible to 12 cluster nodes

12-node refresh disk management can see the 08 storage allocated to it, but it is not added to the 12-cluster storage

Execute the cross-cluster migration wizard and select CSV. You can see that we cannot select one of the virtual machines alone, we can only check the overall CSV.

Click the next step, this step is very critical. You can choose to map the virtual switch connected to the virtual machine, and the adapter connected to the source virtual machine in the future, which is mapped to the switch on the new cluster node. The cross-cluster migration wizard automatically helps us discover this, so we don't have to think about manual operation.

Running the progress bar performs replication of the virtual machine configuration, which is usually fast because it is a light copy and does not include the contents of the virtual machine

After completion, you can see the summary report, indicating the completion results and the follow-up actions to be performed after completion. You can also view the report for more details.

Offline shut down the old cluster virtual machine and CSV, and downtime begins

Unallocate CSV storage to 08 old cluster

12R2 cluster online online CSV

12R2 cluster online virtual machine resources

Click on the virtual machine configuration to see that it has been migrated, and the connection of the network adapter has been mapped in the wizard to ensure that the virtual machine can be connected to the network when it is powered on.

The measured virtual machines can be migrated normally from the nodes in the 12R2 cluster.

At this time, the old cluster has been replaced and the roles and virtual machines can be deleted.

At this point, we have completed the smooth transition of the virtual machine and DHCP role from 2008R2 to 12R2, but here 12R2 still needs to connect to the old storage of 08R2, because the virtual machine uses the CSV above, but we can also use the storage migration on the 12R2 cluster. We also use the storage newly allocated to 12R2 to make CSV. This is an old 08R2CSV, a new 12R2CSV, and if the old storage no longer wants to be used. Can be replaced in this way

There are two CSV,CSV01 in the current cluster. The CSV,VM connected to the old storage is the CSV connected to the new storage by the 12R2.

For the virtual machine to run the storage migration wizard, you can see that in the cluster, storage migration becomes like this. Instead of calling it storage migration, Lao Wang might as well call it a storage copy. Because this wizard is responsible for helping us copy and paste the virtual machine configuration and disk files to another location, and the files in the original location still need to be deleted by ourselves. We need to manually create a folder for the target virtual machine so that all files will be under the folder, otherwise it will remain the default.

Fortunately, we can carry out different storage migration goals for multiple virtual machines in one wizard!

Click launch, Go, to open the door to happiness! After the migration is complete, you can see that the virtual machine storage is already in CSV2, the new storage

Delete the old CSV directly in the cluster because it is no longer useful

Now the old cluster can be completely replaced, the old storage can be completely replaced, we have completely replaced the new cluster, the replacement of new storage!

You can see that although the migration of virtual machines across clusters requires that the new and old clusters are connected to one storage, this is only temporary. When the virtual machines are migrated to the new cluster, we can still migrate the storage of the virtual machines online to the new CSV by adding a new CSV, completely using the new storage and the new cluster.

For the above cross-cluster migration of WSFC from 2008R2 to 2012R2, do you find it very helpful? If you need to know more, please continue to follow our industry information. I'm sure you'll like it.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.