OK, so I feel like I short changed you a bit on my last blog post Part 1 – Configuring Site Recovery Manager (SRM) With HP StoreVirtual VSA as I did’t mention what the heck the Storage Replication Adapter does.
Think of the Storage Replication Adapter as the ‘link’ between your storage vendor’s hardware and SRM. Essentially it allows SRM to peer down into the murky depths of your SAN and issue commands which would otherwise need to be done by the administrator. I think an example is in order, the best one I can think of is when we use SRM to do a test failover (something which we will do later). SRM uses the SRA to allow you to replicate recent changes to the DR site, then it takes a snapshot of the read only replicated volume, changes it to read/write. How cool’s that we just ‘click the button test’.
Right now that’s covered off, let’s jump into vCenter. Launch vCenter and you will see, err nothing different. Why’s that? well we need to install the plug in for VMware vCenter Site Recovery Manager Extension.
Click Plug-Ins from the top menu bar in vCenter and then Manage Plug Ins
Then click on Download and Install under Available Plug-ins
Most likely you will get a Security Warning in relation to your certificates and underlying PKI unless you have a trusted SSL. In my case I don’t so I will tick the box and ignore
As Windows also likes to give us warnings we have another one in relation to running the installation, click Run. Then we get to choose the language (it would be nice to get an English (United Kingdom) for us UK folks) anyway, hit OK.
Click, Next, Next and hit Install, and you should end up with a Finish button.
If all has gone to plan, you should see VMware vCenter Site Recovery Manager Extension with an ‘Enabled’ status
I strongly suggest you do the same at the DR site so you keep everything linear.
Next step is configuring the key components of SRM.
SRM Site Connection
Finally we can click some stuff in vCenter and play with SRM. Click Home and you will see a new Icon under Solutions and Appliances ‘Site Recovery’ I don’t know why but it reminds me of a super hero logo, must be the lightening bolt.
Launch Site Recovery and we are at the landing page, this is where you will spend alot of time.
You will notice that we can only see one site being Production (Local) as we have yet to configure the connection between both vCenters and SRM.
To do this, select Configure Connection from the ‘Commands’ menu on the right hand side
Then enter the address and port of the remote vCenter Server, in this case VMF-ADMIN02 and hit Next
We get another question about certificates, this time we need to validate the vCenter Server Certificate at our DR site, Click OK
We now need to enter the credentials of a user who has rights to access VMF-ADMIN02
Amazing, we have another certificate warning, click OK again. Hopefully, if all goes well, you should see all green ticks and then hit Finish.
Time to authenticate into VMF-ADMIN02, oh by the way, get used to entering your credentials a lot!
Click OK, and ignore the next security warning (I swear VMware is now trying to wind us up). Voila we should now see both site Production (Local) and DR.
You have probably started to click around on the tabs at the top called Resource Mappings, Folder Mappings and Network Mappings. These are logical links between our Production and DR sites, which state if you we use the Port Group LAN in Production when we failover to DR, use Port Group DR LAN in DR. That make sense?
Let’s configure it, I’m sure the penny will drop, if it hasn’t already.
The first tab is Resource Mappings, we are going to configure a mapping between Cluster01 at Production and Cluster02 at DR. To do this we click Cluster 01 and select Configure Mapping.
Expand, Datacenter02 and then select Cluster02 and hit OK
We now have a mapping (logical link) between Cluster01 and Cluster02. Boom!
You can also map Resource Pools between locations, I haven’t created any in this example. One thing to note is that SRM will not create Resource Pool’s in your DR location, you will need to configure and maintain these on a manual basis.
TOP TIP: Don’t forget to do your reverse mappings, DR to Production for failback
Exactly the same principal as Resource Mappings, these create a logical link between our Production and DR sites Folders. I don’t have any folders below, instead I have linked the Datacenter01 and Datacenter02.
This is where things start to get interesting. The network considerations in my opinion are the greatest consideration in your SRM design. Depending on your inter site link, this will have massive implications on what you can or can’t achieve. Let’s discuss these a little further detail below.
Point To Point Link
Let’s say that you have two sites which are connected by Point to Point 100 Mb/s inter site link. You don’t meet the requirements to use a vSphere Metro Storage Cluster, however why would you want to change all of your virtual servers IP address’s in DR? I certainly wouldn’t want to, the risk of third party applications not working as they have been programmed with a static IP.
In this scenario I would recommend getting your network team to stretch VLAN’s between sites and when you failover to use the same Subnet, VLAN and Port Groups.
MPLS/Site to Site VPN
This is the most common scenario and you really have two choices. Either RE IP or not to RE IP. What do you mean Craig, haven’t I got to RE IP as we are on a different site? Well the answer to that is no.
What you can do is get the network team to create the same VLAN at DR as you have in Production, but the key is ensure that the VLAN is shut down! When you failover use the same Subnet, VLAN and Port Groups and then perform a ‘no shut’ on the VLAN and it will work.
TOP TIP: Ensure that you shut down the inter site link port
We are going to make things complicated for ourselves as we are going to RE IP, so that you can see how that works.
Anyway, back on topic when we failover to our DR site, I want my VM’s in DR to be connected to specific Port Groups. These will be:
- LAN > DRLAN
- Backup > DRBackup
We don’t need to worry about our iSCSI connections as SRM combined with the SRA will work that out for us.
If you remember on the first blog post I showed my subnet’s but, as a quick reminder.
- LAN 192.168.37.0/24 VLAN 1 will become DRLAN 192.168.38.0/24 VLAN 51
- Backup 10.37.30.0/24 VLAN 30 will become DRVLAN 10.37.31.0/24 VLAN 31
We follow exactly the same procedure for our Network Mappings as we did for the Resource and Folder Mappings, so I want go over old ground. You should end up with something like this.
What’s this thing called a Placeholder Datastore? What does it do? Well first of all the datastore only needs to be small, 10GB should be fine for all but the largest environment Essentially they contain all of the Virtual Machine configuration files. We are going to RE IP when we perform test failovers, if we changed the IP address of the Production VM in this scenario it would end in dramas.
I have created two Volumes as follows:
- SATAPLACE01 which will be the Datastore for our Production Site
- SSDPLACE01 which will be the Datastore for our DR Site
If you need a hand creating a Volume using HP StoreVirtual VSA, refer to my previous blog series How To Install & Configure HP StoreVirtual VSA on vSpehre 5.1
From these two volumes,I then created two Datastores one at Production called SATAPLACE01 and one at DR called SSDPLACE01.
Back to SRM, let’s click on the Placeholder Datastores Tab and then on Configure Placeholder Datastore
Select SSDPLACE01 (yep SRM uses the opposite sites Datastore to hold VM configuration files, makes sense as when your Production site goes down, SRM knows what VM configuration is needed).
Once done, it should look like this.
Then all we need to do is repeat the process at the DR site.
We are cooking on gas! Time to move onto Array Managers.
Array Manager, what’s that all about? Well it’s exactly what it says it is, SRM uses the Array Manager to guess what, manager the array.
If you remember we installed the HP P4000 Replication Adapter in Part 1 – Configuring Site Recovery Manager (SRM) With HP StoreVirtual VSA so the Array Manager section is where SRM delves into our SAN and discovers which Volumes are being replicated and based around this sees what Volumes are being replicated.
With this in mind, we need to configure SRM to let it know where to find the HP StoreVirtual VSA.
Select Array Managers from the left hand side menu, then ensure you are on Production (Local) and then select Add Array Manager
Give the Array Manager a name, I’m going to call mine ‘Production – StoreVirtual’ and Click next
Now we need to enter the IP Address of the HP StoreVirtual VSA and a username and password with credentials to log into the SAN.
TOP TIP: Enter in the VIP Address of the HP StoreVirtual VSA
Click Next and then hopefully you will see a ‘green success’ tick
Click Finish and repeat the process for the DR site.
Once, done you should see you SAN’s on the left hand side
You will notice that under Array Pairs we don’t have anything listed, why’s that? Well this is the area where SRM shows us the replicated Volumes, and as we don’t have any, nothing appears.
Watch out for Part 3, when we will be replicating Volumes between our Production and DR sites.