Now that I’ve covered replication on the PowerFlex, we can move right into VMware SRM. Like all storage-based replication, PowerFlex has an SRA that integrates with SRM and executes those tasks specific to the PowerFlex. The current version of the SRA is 18.104.22.168 and supports through PowerFlex 3.x. Support for 4.x is currently undergoing certification. Here is the current compatibility matrix:
The noted two footnotes there refer to (1) support for SRM patches, and (2) that equivalency for 8.4-8.6 is granted by VMware for the SRA. This simply means that if your SRA qualified for release 8.3, you automatically get qualification for 8.4 – 8.6.
There is more than enough SRM documentation on this blog, let alone all VMware’s docs, so I’m not going to cover any basics at all. I take for granted that if you’re reading this you are very familiar with VMware SRM. Therefore I’ll only be discussing the PowerFlex SRA. Let’s start with the installation.
PowerFlex SRA Installation
I recently updated the Release Notes for the SRA and re-wrote the installation section so I am going to include it here since it uses my existing environment. As there are a lot of screenshots below so it is easy to follow, I’ve made them smaller. Just click on them to enlarge.
Dell recommends using a custom System name for the PowerFlex cluster rather than the default generated one.
Start by going to the VMware download site and find the PowerFlex SRA.
Extract the Dell EMC PowerFlex SRA file from the downloaded zip file (RN will also be present):
Navigate to the SRM Appliance Management UI at: https://:5480.
The Welcome to VMware Appliance Management dialog box appears.
Enter the credentials (default administrator user is ‘admin’), and select LOG IN.
In the SRM Appliance Management dialog box, navigate to Storage Replication Adapters in the left-hand panel. Next select NEW ADAPTER in the right-hand box.
In the New Adapter dialog box, select UPLOAD and navigate to the file Dell_EMC_PowerFlex_SRA_22.214.171.124.tar.gz
The file will immediately begin uploading.
Before adding the PowerFlex clusters in SRM, an additional configuration step is required. The root certificate of each PowerFlex cluster must be uploaded to its respective SRA. This certificate can be the same one used when replication was configured between the two PowerFlex clusters (previous blog post), or it can be extracted again. The following steps walk through the process. It must be completed for each SRA.
Obtain the root certificate, in PEM format, of the PowerFlex cluster. If there is no generated certificate, it can be exported from the primary MDM instance using CLI. Note that the name of the certificate file is user-defined.
scli --login --username admin --password password scli --extract_root_ca --certificate_file /tmp/root_ca.pem
Log back into the SRM Appliance Management and navigate to Storage Replication Adapters.
Select the three dots in the corner of the Dell EMC PowerFlex SRA Plugin box and choose the menu Download configuration archive. This will initiate a download of the SRA configuration files in the form of the file sra_powerflex~latest.tar.gz.
Using 7-Zip (or another unzip utility) as in this example, extract the archive zipped file and the tar file within it. There will be two files: config.ini and sra-configuration-version.txt
Copy the root certificate of the appropriate PowerFlex cluster into the same directory.
Next, edit the config.ini file with a Notepad or a simple text editor. Do not use Microsoft Word to avoid adding extra characters. Once open, look for the variable ca_path = under the [ssl] section.
Edit the config.ini by typing in the name of the root certificate then saving the file.
The three files can now be re-packed. The following walks through the steps using 7-Zip.
Right-click on the files and select 7-Zip -> Add to archive…
Provide a name for the file with .tar suffix and set the Archive format to tar.
Next, right-click on the newly created tar file, in this example dsib2227.tar, and again choose
Add to archive…
Provide a name for the file with .tar.gz suffix and set the Archive format to gzip.
Now upload the file to the SRA. Using the same process as downloading, select the three buttons and choose Upload configuration archive.
Provide the file and select UPLOAD. Another green box will appear indicating success.
With the SRAs installed, you can add the array managers. The array managers rely on the Gateway server that is part of each PowerFlex cluster. This is the same interface used to do the installation.
In SRM, navigate to Configure -> Array Pairs and select ADD.
Select the PowerFlex SRA in the next screen.
Start with configuring the Local array manager. Provide a site name, and then the IP of the Gateway on the first cluster. PowerFlex supports bi-directional replication, so if you plan on replicating both ways, it really doesn’t matter which Gateway IP you start with. If you have a lot of volumes on your PowerFlex and all of them are not part of VMware, you can set a prefix if the VMware volumes have a format. I’m allowing full discovery below.
Next, configure the remote array in the same manner.
An initial device discovery is automatically executed, and if you have pairs configured, as I do, they will display. Note I have bi-directional volumes.
The PowerFlex SRA supports all SRM operations: test failover, planned migration, failover and reprotect. It does this by utilizing the same capabilities available to the user through the GUI. The menu options in the Remote Consistency Group (RCG) screen provide both the test failover and failover capabilities. Below in the green box I have the two commands that are used for SRM test failover: Test Failover (SRM run test failover) and Test Failover Stop (SRM run test failover cleanup). In the red box are the commands for SRM failover: Failover (SRM planned migration or failover – remember they are essentially the same) and Reverse (SRM reprotect).
Now that I’ve completed the configuration, along with the previous setup from my replication post, I am going to narrate a demo of running SRM. For reference, here is some detail of the environment in the demo. I’m including it because, in particular, there are a number of different names used in the PowerFlex setups which are similar but have different functions.
There are two PowerFlex clusters:
When you pair these clusters for replication, you must assign a name to each one. Now while it could be the same name as the cluster itself, that can be confusing so I’ve named them similar but obviously different:
- Replication Peer Name: PF_Protection
- Replication Peer Name: PF_Recovery
After pairing, I created a Replication Consistency Group (RCG). This is the same one used in the previous post. It is Test_RCG. It has a single device pair that is 112 GB (expanded from the original 8 GB). The naming used in the RCG screen adds even more confusion, because PowerFlex only uses the Peer Name above at each local site. So if I’m on my Protection Cluster, my local replication peer is called PF_Protection, but my remote peer is called Recovery_Cluster. This makes absolutely no sense to me, but here is what I mean:
Anyway, the final component is VMware. So my two vCenter clusters:
- Protection Site: Boston_Cluster
- Recovery Site: London_Cluster
That should be good. I don’t go too fast in the demo and obviously you can pause it if you need to check the above info. In the demo I’ll run through the basic PowerFlex setup, creation of a protection group/recovery plan, and then run test failover, failover, reprotect. That will give us the range of capabilities. Because of the tasks and explanations the video is about 15 minutes, hence why I didn’t want to try to write this up. I recommend watching the video on YouTube at 720p because below is fuzzy if expanded.