SRM with VMAX 20K Part IX: Provisioning SRDF/A replicated volumes

In this post I am going to run through setting up replicated volumes on the VMAX 20K. There is a little assumed knowledge with this, I only know what I’ve learnt through reading up and dabbling with these monster arrays. As discussed in Part I of this series, there is already a VMAX Configured with SRDF connectivity enabled. I’m not going to go into much detail about this purely because it would require a lot more than a single blog post. I’m mostly concerned with getting replicated volumes ready in a state for SRM.

To give a very brief description of some of the terminology being used:

R1 Device / RDF1 Device Group = Source (Read/Write) volume or group of volumes on Primary Site Array.

R2 Device / RDF2 Device Group = Target (Write Disabled) volume or group of volumes on Secondary Site Array.

SRDF Group = Used to define a relationship between two Symmetrix Storage Arrays that are connected by SRDF. It defines SRDF Director ports assigned on local and remote arrays. Each SRDF volume must be assigned to an SRDF Group as an R1 or R2 device.

Device Group = User defined object which groups together Symmetrix devices. It enables you to call multiple R1/2 volumes by giving them a group membership. Device Groups propagate using GNS (Global Name Service) which copies group information to remote SYAMPI configuration databases to enable consistent naming schemes. (In my head I think of AD Groups and replication between DC’s to give it a simile).

There is plenty of information available out there from EMC and affiliated groups of people, as I mentioned in Part I of my series, which have enabled me to learn and get this working. There is a lot of documentation and some of it is heavy going if you aren’t an EMC SAN guru, but it is well worth reading.

Right, on with the show!

Configure Volumes

The process of configuring a volume on the VMAX is quite straight forward. I use the Unisphere Web Console because I find it easy and quite idiot proof.

1) Login to each site VMAX Array and configure a new volume as shown below. Ensure under Advanced Settings that the device has a dynamic capability of “RDF1_OR_RDF2_Capable”.
SRDF_1In this example, I needed 4 volumes on each site.

2) The volumes should create as normal and just be normal TDEV (Thin Device) volumes at this stage.

Primary Site:
SRDF_2

Secondary Site:
SRDF_3

3) At this time I recorded all volume numbers, WWN’s, pool memberships, etc for quick reference and good practice on documentation.

Create SRDF Group

1) On the Primary Site VMAX, navigate to Data Protection, Replication Groups ans Pools, SRDF Groups:
SRDF_4

2) Select “Create” and then fill out information on the group. Including a: Group Label (Name), Communication Protocol, Unique Group Number and Director ports that you have assigned for SRDF on the VMAX:
SRDF_5

3) Close the success message. A new group label should appear in the Web UI.
SRDF_6

4) It is also possible to check the group via the SYMCLI. This can be done on the Array Manager server as configured in Part V.

SRDF_7

SRDF_8 Where XX2 is the last 3 digits of your Symmetrix Array as seen in step 4. For me, XX2 is Local (R1) and XX4 is Remote (R2).

5) Create a text file in a directory on the Array Manager host. List the volumes created earlier from the Local and Remote arrays. As below:
SRDF_9

6) Run CMD as Administrator and enter the following command, when prompted hit “Y” to confirm:

SRDF_10

SRDF_11
This command pairs the devices listed in the txt file and creates an RDF1 group on the local array on SRDF Group 4 (Created above) in Adaptive Copy mode.

7) Check the status of the SRDF Group and the paired volumes within, using the device pairs file:

SRDF_12
Notice that the pair state is suspended. The astute reader might also see that I have 5 volume pairs in this group rather than 4 as per my earlier creation. This was because I needed another later on so I added it in before recording my steps.

8) Now the devices have been paired, start the non-consistent SYNC of the devices. Run the command below and accept the prompt with “Y”:

SRDF_13

9) Check the status of the SYNC. In this case, both volumes are “empty” and thin provisioned. Run the command:

SRDF_14

The state “SyncInProg” can be seen. This command refreshes every 5 seconds showing the progress of the volume synchronisation.

10) Once complete, run a final check on the status of the SRDF Group:

SRDF_15

11) As the SRDF SRA required Sync/Async replication, set the replication mode of the SRDF Group to Asynchronous (SRDF/A), confirming the prompt:

SRDF_16

SRDF_17

Configure Device Group

This will enable the calling of volume pairs within an SRDF Group within having to use a text file to specify the ID’s. This will replicate across both arrays using GNS. I create and configured the device group on the Primary array.

1) Create the device group and give it a name:

SRDF_18

2) Add each of the devices to the group and check it has worked by querying it:

SRDF_19

Repeat with all devices that are R1. The group will automatically create and populate on the Remote Array.

3) Run a check on the new Device Group by showing the full details:

SRDF_20

SRDF_21

If you check the status of the output above, you will see that the device group is not consistent. Consistency groups are required for SRM.
SRDF_22

4) Enable device consistency on the group, run the following command and accept the prompt:

SRDF_23

After this, re-run the same show command on the group and check the consistency state
SRDF_24

5) It is now possible to query the Device Group and it’s contents without having to use the manual text file.

SRDF_25

6) The volumes can now be presented to the VMware ESXi storage groups that are configured for you hosts on each VMAX. Present the R1 devices to your Primary Array Storage Group and the R2 devices to the Secondary Array Storage Group:
SRDF_26

Conclusion

This finishes off quite a large post containing a bit of a crash course in configuring replicated volumes on a VMAX 20K using SRDF/A.

I have experienced quite a few storage arrays in my time but the VMAX is on the high end of in terms of capability, performance and availability. The above, for me, was a challenge in understanding some concepts and getting it to work. I had a few difficulties along the way which EMC Support were very good with assisting me on. I’m always happy to bash support when they are useless but in this case I got through to a couple of guys in Ireland who were excellent.

As I mentioned at the start, there is a wealth of information that could have gone into this. I’ve covered the main bits in terms of implementation, but there is a fair bit more work around configuration and general administration of the VMAX that might not be here. Although, I’m happy to ask questions or deflect to the right person if anyone is reading this and has any further questions.

In the next post, I’m going back into VMware and configuring the environment ready for SRM.

4 comments

    • Ryan Harris
      Ryan Harris

      Hi Geoff,
      Thanks for the comment! Unfortunately, I’m no longer working with VMAX Arrays so I won’t be able to cover SRDF/S. There is quite a wealth of documentation out there though, it might just take a bit more going through than the cut down version I posted. Cody Hosterman covered quite a lot in his documentation I posted in the first of my series. Might be worth taking a look!
      Ryan

Leave a Reply

Your email address will not be published. Required fields are marked *