Maintenance procedure for low-impact changes

From ImageWeb

Low-impact change

Low-impact maintenance procedure is intended to be invisible to the users of the system. But there is a small risk that such a change affects functionality of the system in a way that impacts users of the system.

Hence, we propose the following procedure for changes judged to be low-impact.

Standard Procedures

Since ADMIRAL uses separate instances of VMs for different groups of users, it is essential to roll out the changes in a standard manner.

• The change first needs to be tested on the local test system and then the IBRG. Ensure that all the tests pass.

• To roll-out the changes to all the groups, the Research Group Leader of each group needs to be notified of the forthcoming maintenance about 2-5 days ahead (preferably at the supper end of that range) with the following information:
  • The date/time proposed for performing the system maintenance and the duration of the maintenance. (Ensure to include the testing time after roll over of changes.)
  • A brief description of the nature of the maintenance
  • Informing them that the maintenance is not expected to have any visible impact on the operation of their ADMIRAL system
  • Informing them that there is a small risk of disruption during the maintenance
  • Asking them to contact the person in-charge of the maintenance to re-schedule the maintenance if they are intending to be doing any urgent work at the time

• After rolling out the changes to each group, run all the tests against each and ensure they pass before making any changes to the next group.

• Clean up any test-users/data created as part of the system testing before sending a final email to notify that the maintenance is over.

• Notify the Research Group Leaders about the completion of the maintenance.