|
|
|
 |
| Predictable Change Schedule |
|
Data center systems* require periodic service and upgrades to operate at their highest capability and efficiency. To minimize system downtime and promote predictable update times, the SDC has set regular maintenance windows for all systems.
A maintenance window is “a defined time frame in which the performance of planned maintenance work has the least impact on customers or the enterprise.”
*Including servers, network equipment, mainframes, etc
|
|
|
| Using a Maintenance Window |
|
Change Management
The SDC notifies it customers (through the SDC change management process) of the intent to perform work on systems. In most cases, updates performed 1) do not take the full amount of time allocated on a given day nor 2) is every day used for updates. Some work performed during the maintenance window is recurring and is pre-arranged through an approved standing change request. An example is regular server reboots.
*Exceptions
Business Requirements
Some systems require updates outside the standard window due to the nature of the certain state businesses. These systems are handled as an exception and work re-scheduled through the SDC change management process. There are also instances where updates are performed prior to the next maintenance window scheduled. In these instances, urgent or emergency SDC change management processes are followed.
Notification Process
- Here’s how the agency requests SDC not to use a change window: Ten business days or more prior to the change window an agency requests the exception; e.g., an agency's year-end business process is adversely affected by any work performed in a change window.
- The agency CIO sends the SDC a written request not to use a particular maintenance window during a certain event.
|
|
|
| Maintenance Window Timetable |
|
|
Service Area
|
Frequency
|
Start
|
End
|
Comments
|
Mainframe
|
|
|
|
See below
|
DAS
|
1st Sunday
|
6:30 a.m.
|
8:30 a.m.
|
|
DHS
|
2nd Sunday
|
6 p.m.
|
8 p.m.
|
|
ODOT
|
3rd Monday
|
3:30 a.m.
|
5:30 a.m.
|
|
|
|
|
|
|
Computing
UNIX
|
Monday - Thurs
|
10 p.m.
|
2 a.m.
|
See below
|
|
Sunday
|
6 p.m.
|
2 a.m.
|
|
|
|
|
|
|
Computing
ISeries
|
Saturday
|
6 p.m.
|
Midnight
|
See below
|
|
|
|
|
|
Computing Distributed Systems
|
Monday - Thurs
|
10 p.m.
|
2 a.m.
|
See below
|
|
Sunday
|
6 p.m.
|
Midnight
|
|
|
|
|
|
|
Storage
|
Monday - Friday
|
6 p.m.
|
10 p.m.
|
See below
|
|
Sunday
|
6 p.m.
|
Midnight
|
|
|
|
|
|
|
Network & Security
|
Sundays
|
6 a.m.
|
8 a.m.
|
See below
|
|
|
|
|
|
|
|
|
| Maintenance Window Details |
|
Mainframe
Scheduled for regularly occurring, preventive mainframe IPL’s as shown above.
- The IPL’s begins at the start of the window and takes about 45 minutes to complete.
- Patches and other changes are also scheduled.
- Typically patches are installed prior to IPLs and activated as part of the IPL process.
|
 |
Computing - UNIX
Scheduled regularly for server patches, updates, and reboots.
- The regular reboots start in a staggered sequence and take about two hours.
- Not every server is rebooted each week
- Reboots are specified on a detailed schedule and communicated to the agency as part of a standing change request.
Servers are constantly monitored. SDC Operations notifies if rebooted servers are not functioning properly.
|
Computing - iSeries (AS/400)
Scheduled Saturday nights, from 6 p.m. to midnight.
- Systems administration activities that remove or restrict customer access to an iSeries LPAR are scheduled during this window.
- Each iSeries agency has an LPAR save schedule, which does not match the SDC iSeries maintenance window. Each agency has a system save paradigm that includes a complete LPAR save each week (an IPL—POR—which applies PTFs if present) and periodically saves throughout the week.
Storage
Virtual Tape System (VTS), Automated Tape Library (ATL) and Storage Area Network (SAN) – Scheduled regularly to perform vendor recommended updates. Redundancy is built into the system to reduce or eliminate impact to customers.
Network
Scheduled regularly to perform network updates, apply security upgrades and fixes, and replace equipment. For core and distribution infrastructure, redundancy is built into the system so any disruption to service is minimal.
Security
Scheduled for redundant systems, specifically firewalls. The impact of routine maintenance does not typically affect SDC customers. Security systems involved include updates; e.g., firewalls and monitoring (MARS).
|
|
|
| Maintenance Activity Types |
|
Planned Maintenance
Any hardware or software change causing a service disruption or degradation to service scheduled:
- in advance through the change management process and
- occur during the standard maintenance window.
Work occurs within the scheduled maintenance window and through normal change management procedures.
Urgent Maintenance
Any hardware or software change causing a service disruption or degradation to service where:
- maintenance performed outside of a normal maintenance window and
- short notice customer notification must occur.
Work occurs before the next scheduled maintenance window and outside the normal change request processing schedule.
Emergency Maintenance
Any hardware or software change performed with little or no notice, usually initiated by a system outage or security incident. Work occurs immediately and outside the normal change request processing schedule with an after-the-fact change notification.
|
|
|
|
