How do others maintain their Data Center Disaster Recovery (DC DR) SAS environment?

[MG18]

We have a SAS deployment at our DR site.

My question is,

How do others manage the installation of hotfixes/updates to their DR environments ?

How do synchronisation happens between Production and DC SAS Server (Meta , Mid tier, Compute , SAS VIYA server)?

Is it happens automatically ? What technology others use for the synchronisation ?

The Production and DR SAS environments share the different network alias (DNS NAME , hostname ), so both environment can be running at any one time. 

what is the others reverse synchronising strategy i.e.  When running on DR and switching back to production ? 

Do I need different licenses for DR servers ? 

Many thanks in advance 🙂 .

[Kurt_Bremser]

Since we run our data warehouse in a virtual machine (AIX LPAR), we do not need to manually synchronize anything. The root volume group is backed up daily, and in case of a disaster, the image is started up in the secondary location. Data is mirrored across the sites synchronously by the SAN infrastructure.

Since the SAS software is installed on the mirrored volumes, all updates propagate automatically.

[JuanS_OCS]

Hello @MG18 ,

gerat topic, although really broad, you would need proper architecture sessions to determine.

You can hear many stories here, which can be great for their specific environments, but not necessarily can be applied to your environments.

In any case, ideally, your DR site should be updated every time your main site is updated, following D-T-A-P-DR procedures of promotion. The data of course should be in sync all the time, as frequent as possible, while ot blocking/locking your Production data and minimizing performance reduction.

In terms of technologies, it matters a lot the type of machines and infra you have in place: physical machines? VMs? In any Cloud? 

You will always need licenses, legally speaking, for different servers. The costs is something to discuss and negotiate with your SAS Account Manager/Representative, and the type of DR site and method (cold, standby, hot) will influence the total costs.

This being said, keep in mind that all is part of the DR plan, and insights from one topic, will determine changes in another topic. I.e: license for hot DR is too expensive, I want to be cost-savvy, therefore I need to change my DR design to implement additional procedures and have a cold or standby DR, and each of these 2 will have different costs, different design, different procedures and policies.

I hope it helps,

Best regards,

Juan

[Anand_V]

In my experience usually if the SAS servers are part of VM farm then there are available technologies that can sync everything from existing VM to DR VM automatically. It is usually scheduled from the VM support team managing the infra. Since there is a requirement of hostnames should be same for SAS if that cannot be achieved seamlessly then you can use DNS or alias name and do a switch while moving to DR.

If this installation is on physical machines then you have storage replication technologies that can perform the job. You should talk to your Infra team on how are they doing it for other applications.

If you choose for a hot-hot approach I believe you will have to pay license fees for both the SAS environments. Again It's a good discussion to have with your account manager.

[Jackie_sas]

You might be interested in this SAS Global Forum paper:  "Do You Have a Disaster Recovery Plan for Your SAS Infrastructure?"  https://support.sas.com/resources/papers/proceedings17/SAS0565-2017.pdf   It provides background, SAS specific references, and a few examples.