There may be occasions when you want to shut down the entire fabric, not just particular partitions or platforms. The following procedure describes how to do this. You must shut down the fabric platform by platform.
If you want to shut down just parts of the fabric, not the entire fabric, you can perform just a subset of the procedure.
Typically you first stop your applications in whatever manner is correct for them. For example, this could consist of stopping web services, followed by stopping application logic, followed by stopping database services.
It is recommended that you take a back up of an enterprise partition platform’s configuration settings before powering down the platform.
Power down the first enterprise partition platform, using the Soft Shutdown command from the Fabric Manager. Doing this accomplishes the following:
Performs an implicit graceful shutdown of all of the partitions on the platform (and the partition operating systems)
Implicitly stops s-Par
See Performing Soft Shutdown on a Platform for the procedure.
This removes DC power to the EPP. Note that this does not remove AC power from the EPP’s platform management card, which is powered by keep-alive power from the system’s uninterruptible power supply (if one exists).
Approximate duration of step: Typically three to four minutes. It is recommended that you wait ten minutes before taking remedial action.
If step does not complete successfully: If a given operating system does not respond, do as follows.
Individually stop the offending operating system, for example, using Remote Desktop.
If this does not work, do a Force Halt (unorderly stop) of the partition, using the Fabric Manager.
Do a Soft Shutdown of the whole platform again, using the Fabric Manager.
If the preceding step does not work, then use the Fabric Manager Power Off command to do an unorderly power off of the platform.
Power down all of the other enterprise partition platforms.
Perform an orderly shutdown of the Linux operating system on the Fabric Management Platform, using one of the following methods:
Note: If your system is configured with two FMPs for high availability, power off the FMP that is associated with the Slave status, and then power off the FMP that is associated with the Master status. Use the FFM High Availability Status pane to note the Status field value for both FMPs.
From a putty terminal session, specify shutdown –h now.
After logging in as the root user, perform a poweroff command.
Doing this stops the Fabric Manager software, stops the Fabric Management Platform’s Linux operating system, and powers down the Fabric Management Platform.
If step does not complete successfully: Physically press the power button on the Fabric Management Platform.