« The practical and the prototypes | Main | Enter a line break in a cell (Excel) »

Moving on Up(time)

As you've probably heard, SCLS Automation will soon be moving to new and improved quarters shared with SCLS Administration. Moving with us is the LINK Network Operations Center (NOC), a.k.a. "the computer room". This includes Dynix, LINKcat Web, LibraryOnline, and all the SCLS network services that member libraries and their patrons rely on every day.

The bad news is that there will be a few unavoidable interruptions of service during this transition. The good news is that we should be able to keep disruptions to a minimum. We'll be moving "on uptime" to the extent possible.

How Does That Work?

SCLS network services are provided by a large number of servers, routers, switches, cables, disks, databases, operating systems and software applications. These pieces are all layered on top of one another in complex dependencies. That tends to make this move difficult. However, many services are built up from modular interchangeable parts and that makes things a whole lot easier.

Imagine that network services are provided instead by three library clerks. One can be working the circulation desk, one doing shelving, and one taking a break at any given time. They can trade roles at any point such that services are provided continuously even while there is some downtime.

Now imagine that these network clerks are identical triplets who always dress alike; if you're not looking right at them when they trade roles, you won't even know it happened! So, the real trick for moving on uptime will be to ship individual pieces to the new building while they are "on break". Then we'll do component role swapping early in the morning and late at night, when you're "not looking". Real downtime for individual "clerks" will be extensive, but library staff and patrons will not perceive most of it.

What's the Downside?

Unfortunately some service roles can only be filled by specialized components. Technically these are still interchangeable parts, but they're also expensive. SCLS simply can't duplicate some systems. The Dynix database server is the chief example; the only way to move it is to turn it off, extract it, package it ever so carefully and put it on a truck.

Because we need professional movers and SirsiDynix support, the Dynix move will happen during business hours. You should expect Dynix (and LINKcat and TeleCirc) to be down most or all of Friday, November 20th. Some additional Dynix downtime will occur the following Monday morning when a Sun technician inspects the system to certify our Sun maintenance agreement.

On that Friday and Monday we expect each day's patron notices to be printed and put in delivery as normal. All other network services will remain up during the Dynix move and certification downtimes.

Other specialized, unique systems include LibraryOnline and the SCLS email server. Required downtime for these moves should occur outside of library business hours, but you might notice it if you're a night owl or an early bird. These moves have not yet been scheduled. For these and all other service issues, stay tuned to SCLS mailing lists for scheduling announcements and to the SCLS Status Wiki for live progress reports.

This message from the Bureau of Backup's Supernumerary Systems Branch at the Department of Redundancy was brought to you by the Pleonastic PR Division of the Alternative Options Agency.

Comments

Feed You can follow this conversation by subscribing to the comment feed for this post.

The comments to this entry are closed.