Are you still postponing that upgrade because there just never seems to be a good time to have a planned outage that everyone agrees on? Planning the maintenance of your agents no longer means negotiating rare idle time windows with your colleagues. Often your agents are running business-critical applications 24X7. You want to ensure scheduling continuity while minimizing upgrade efforts. If an upgrade is not planned carefully, you risk prolonged scheduling downtime and unmet service level agreements.
With IBM Workload Scheduler V9.4 you can upgrade both fault-tolerant and dynamic agent instances with little or no impact on your job scheduling activities. The agents are stopped for the shortest time necessary, and any jobs already running when the upgrade process begins, continue to run as planned, but no new jobs begin execution during this time. Once the upgrade is complete, the agent is restarted and quickly reconnects with its jobs. Any jobs that were actively running before the upgrade that have not yet completed, continue to run, and any jobs that successfully finished during the upgrade procedure report a successful job status.
The same behavior is applied to active agent command-line interfaces and processes such as conman or composer. After the agent upgrade, all processes and command-line interfaces resume and continue to run regularly.
It’s very easy and supported by both agent upgrade methods: the twsinst script that you run on each individual agent workstation, and the centralized agent update that upgrades multiple agents at a time. And if something goes wrong? An automatic backup and restore feature rolls back the agent to the prior version before a problem creates havoc in your environment.