A manufacturing client recently had a close call with their server. They were lucky and it was possible to get the machine back running without having to restore from their virtual server back-up. By design, not luck, we had those backups in two forms, one on and one off site.
The tech was very upset all the same. As he put it to me, restoring from backups after having lost their only server would take a couple days at least, and their operation would suffer if down for a couple hours. Everything was on this one machine.
The list of “everything” meant the clerical and engineering files, the design files, the library of manufacturing machine programs, the job tracking and the accounting data and software. There were some computer functions that could be done without the server, but basic operations for the company as a whole would be held up if this server machine was not available.
So I called the client up and told him what our tech had told me.
“What can we do?” was the client’s obvious and fair question.
I described to the client some of the options available on modern servers. Many of our client’s servers today are “hosts” meaning that they host virtual images of servers. The server one sees on the network is kind of a recording of a server; that is the image.
The physical machine, the host, is not as expensive as it once was, and the economics of owning a couple of them makes a lot more sense today. Then one can back up the entire virtual image somewhere and have a computer ready to put it right back to work on the network within a couple hours. More expensive setups can have an uninterrupted failover, but that is not usually in the small office budget.
The client had a business decision to make. A two hour recovery time vs a two day wait to be back up and running. The difference was one more physical server and a virtual image backup.
“Do it” he told me.
