topic How do others maintain their DR SAS environment? in Administration and Deployment

How do others maintain their DR SAS environment?

matthewmeserve — Wed, 15 Jul 2020 23:06:26 GMT

We have a SAS deployment at our DR site. It is not licensed since it's only use is in the event of a disaster. I'm told by SAS that, once a disaster occurs, they will issue a temporary license for the environment so it may be used until production is back up. My question is, how do others manage the installation of hotfixes/updates to their DR environments without, at least, a temporary license so that SAS could be launched and tested briefly post hotfix installation? Are others simply leveraging the built in OQ/IQ utilities and calling it good?

Curious to know.

Thanks!

Re: How do others maintain their DR SAS environment?

SASKiwi — Thu, 16 Jul 2020 00:11:50 GMT

We also have a DR SAS environment. In our case it is hosted in a separate data centre, 800 km from the Production one, and is run on a warm standby basis. That is it is switched on, but no SAS processes are running. We have an overnight replication process which copies Prod SAS metadata content and SAS data to DR.

The Production and DR SAS environments share the same network alias (CNAME), so only one environment can be running at any one time. IT just switch the IP addresses to move from Prod to DR or vice versa. Under the terms of our licence agreement we only pay for one SAS environment and the same SAS licence runs on both Prod and DR. When running on DR, we get warnings that SAS is not licensed to run on this CPU but everything still works. We don't need a temporary licence.

Whether you do or do not need a temporary licence for your DR environment would depend on the terms and conditions of your SAS licence agreement.

Re: How do others maintain their DR SAS environment?

Kurt_Bremser — Thu, 16 Jul 2020 09:25:34 GMT

Since our SAS server is virtualized, DR is done by booting the image in the backup location.

In the main datacenter, there are two physically separated compartments (firewall between them) with identical hardware, and data (including the image) being mirrored synchronously. In a further, remote location, hardware is kept in standby, and copies of back-upped data are made there daily. If a massive disaster hits datacenter 1 (eliminating both compartments), the remote DC is booted up with yesterday's state. Since all identification (server name etc) is part of the image, SAS does not complain.