Some modifications to a service profile or to an updating service profile template can be disruptive and require a reboot of the server. You can, however, configure deferred deployment to control when those disruptive configuration changes are implemented. For example, you can choose to deploy the service profile changes immediately or have them deployed during a specified maintenance window. You can also choose whether or not a service profile deployment requires explicit user acknowledgement.
Deferred deployment is available for all configuration changes that occur through the association of a service profile with a server. These configuration changes can be prompted by a change to a service profile, to a policy that is included in a service profile, or to an updating service profile template. For example, you can defer the upgrade and activation of firmware through host firmware packages and management firmware packages, such as server BIOS, RAID controller, host HBA, and network adapters. However, you cannot defer the direct deployment of firmware images for components that do not use either of the firmware packages, such as Cisco UCS Manager, fabric interconnects, and I/O modules.
Deferred deployment is not available for the following actions which require the reboot of a server:
Initial association of a service profile with a server
Final disassociation of a service profile from a server, without associating the service profile with a different server
Decommissioning a server
Reacknowledging a server
Resetting a server
If you want to defer the deployment of service profile changes, you must configure one or more maintenance policies and configure each service profile with a maintenance policy. If you want to define the time period when the deployment should occur, you also need to create at least one schedule with one or more recurring occurrences or one time occurrences, and include that schedule in a maintenance policy.
A schedule contains a set of occurrences. These occurrences can be one time only or can recur at a specified time and day each week. The options defined in the occurrence, such as the duration of the occurrence or the maximum number of tasks to be run, determine whether a service profile change is deployed. For example, if a change cannot be deployed during a given maintenance window because the maximum duration or number of tasks has been reached, that deployment is carried over to the next maintenance window.
Each schedule checks periodically to see whether the Cisco UCS instance has entered one or more maintenance windows. If it has, the schedule executes the deployments that are eligible according to the constraints specified in the maintenance policy
A schedule contains one or more occurrences, which determine the maintenance windows associated with that schedule. An occurrence can be one of the following:
One Time Occurrence
One time occurrences define a single maintenance window. These windows continue until the maximum duration of the window or the maximum number of tasks that can be run in the window has been reached.
Recurring Occurrence
Recurring occurrences define a series of maintenance windows. These windows continue until the maximum number of tasks or the end of the day specified in the occurrence has been reached.
Maintenance Policy
A maintenance policy determines how Cisco UCS Manager reacts when a change that requires a server reboot is made to a service profile associated with a server or to an updating service profile bound to one or more service profiles.
The maintenance policy specifies how Cisco UCS Manager deploys the service profile changes. The deployment can occur in one of the following ways:
Immediately
When acknowledged by a user with admin privileges
Automatically at the time specified in the schedule
If the maintenance policy is configured to deploy the change during a scheduled maintenance window, the policy must include a valid schedule. The schedule deploys the changes in the first available maintenance window.
Pending Activities
If you configure deferred deployment in a Cisco UCS instance, Cisco UCS Manager enables you to view all pending activities. You can see activities that are waiting for user acknowledgement and those that have been scheduled.
If a Cisco UCS instance has pending activities, Cisco UCS Manager GUI notifies users with admin privileges when they log in.
Cisco UCS Manager displays information about all pending activities, including the following:
Name of the service profile to be deployed and associated with a server
Server affected by the deployment
Disruption caused by the deployment
Change performed by the deployment
Note
You cannot specify the maintenance window in which a specific pending activity is applied to the server. The maintenance window depends upon how many activities are pending and which maintenance policy is assigned to the service profile. However, any user with admin privileges can manually initiate a pending activity and reboot the server immediately, whether it is waiting for user acknowledgement or for a maintenance window.
Guidelines and Limitations for Deferred Deployment
Cannot Undo All Changes to Service Profiles or Service Profile Templates
If you cancel a pending change, Cisco UCS Manager attempts to roll back the change without rebooting the server. However, for complex changes, Cisco UCS Manager may have to reboot the server a second time to roll back the change. For example, if you delete a vNIC, Cisco UCS Manager reboots the server according to the maintenance policy included in the service profile. You cannot cancel this reboot and change, even if you restore the original vNIC in the service profile. Instead, Cisco UCS Manager schedules a second deployment and reboot of the server.
Association of Service Profile Can Exceed Boundaries of Maintenance Window
After Cisco UCS Manager begins the association of the service profile, the scheduler and maintenance policy do not have any control over the procedure. If the service profile association does not complete within the allotted maintenance window, the process continues until it is completed. For example, this can occur if the association does not complete in time because of retried stages or other issues.
Cannot Specify Order of Pending Activities
Scheduled deployments run in parallel and independently. You cannot specify the order in which the deployments occur. You also cannot make the deployment of one service profile change dependent upon the completion of another.
Cannot Perform Partial Deployment of Pending Activity
Cisco UCS Manager applies all changes made to a service profile in the scheduled maintenance window. You cannot make several changes to a service profile at the same time and then have those changes be spread across several maintenance windows. When Cisco UCS Manager deploys the service profile changes, it updates the service profile to match the most recent configuration in the database.
Configuring Schedules
Creating a Schedule
Procedure
Command or Action
Purpose
Step 1
UCS-A# scope system
Enters system mode.
Step 2
UCS-A /system # create schedulersched-name
Creates a scheduler and enters scheduler mode.
Step 3
UCS-A /system/scheduler # commit-buffer
Commits the transaction to the system configuration.
The following example creates a scheduler called maintenancesched and commits the transaction:
UCS-A /system/scheduler/one-time # set datemonthday-of-monthyearhourminute
Sets the date and time this occurrence should run.
Step 5
UCS-A /system/scheduler/one-time # set concur-tasks{unlimited | max-num-concur-tasks
(Optional)
Sets the maximum number of tasks that can run concurrently during this occurrence.
If the maximum number of tasks is reached, the scheduler waits for the amount of time set in the minimum interval property before scheduling new tasks.
Step 6
UCS-A /system/scheduler/one-time # set max-duration{none | num-of-daysnum-of-hoursnum-of-minutesnum-of-seconds}
(Optional)
Sets the maximum length of time that this schedule occurrence can run. Cisco UCS completes as many scheduled tasks as possible within the specified time.
Step 7
UCS-A /system/scheduler/one-time # set min-interval{none | num-of-daysnum-of-hoursnum-of-minutesnum-of-seconds}
(Optional)
Sets the minimum length of time that the system should wait before starting a new task.
Step 8
UCS-A /system/scheduler/one-time # set proc-cap{unlimited | max-num-of-tasks}
(Optional)
Sets the maximum number of scheduled tasks that can be run during this occurrence.
Step 9
UCS-A /system/scheduler/one-time # commit-buffer
Commits the transaction to the system configuration.
The following example creates a one time occurrence called onetimemaint for a scheduler called maintsched, sets the maximum number of concurrent tasks to 5, sets the start date to April 1, 2011 at 11:00, and commits the transaction:
UCS-A# scope system
UCS-A /system # scope scheduler maintsched
UCS-A /system/scheduler # create occurrence one-time onetimemaint
UCS-A /system/scheduler/one-time* # set date apr 1 2011 11 00
UCS-A /system/scheduler/one-time* # set concur-tasks 5
UCS-A /system/scheduler/one-time* # commit-buffer
UCS-A /system/scheduler/one-time #
Specifies the hour at which this occurrence starts.
Step 6
UCS-A /system/scheduler/recurring # set minuteminute
(Optional)
Specifies the minute at which this occurrence starts.
Step 7
UCS-A /system/scheduler/recurring # set concur-tasks{unlimited | max-num-concur-tasks
(Optional)
Sets the maximum number of tasks that can run concurrently during this occurrence.
If the maximum number of tasks is reached, the scheduler waits for the amount of time set in the minimum interval property before scheduling new tasks.
Step 8
UCS-A /system/scheduler/recurring # set max-duration{none | num-of-hoursnum-of-minutesnum-of-seconds}
(Optional)
Sets the maximum length of time that this schedule occurrence can run. Cisco UCS completes as many scheduled tasks as possible within the specified time.
Step 9
UCS-A /system/scheduler/recurring # set min-interval{none | num-of-daysnum-of-hoursnum-of-minutesnum-of-seconds}
(Optional)
Sets the minimum length of time that the system should wait before starting a new task.
Step 10
UCS-A /system/scheduler/recurring # set proc-cap{unlimited | max-num-of-tasks}
(Optional)
Sets the maximum number of scheduled tasks that can be run during this occurrence.
Step 11
UCS-A /system/scheduler/recurring # commit-buffer
Commits the transaction to the system configuration.
The following example creates a recurring occurrence called recurringmaint for a scheduler called maintsched, sets the maximum number of concurrent tasks to 5, sets the day this occurrence will run to even days, sets the time it will start to 11:05, and commits the transaction:
UCS-A# scope system
UCS-A /system # scope scheduler maintsched
UCS-A /system/scheduler # create occurrence recurring recurringmaint
UCS-A /system/scheduler/recurring* # set day even-day
UCS-A /system/scheduler/recurring* # set hour 11
UCS-A /system/scheduler/recurring* # set minute 5
UCS-A /system/scheduler/recurring* # set concur-tasks 5
UCS-A /system/scheduler/recurring* # commit-buffer
UCS-A /system/scheduler/recurring #
Deleting a One Time Occurrence from a Schedule
If this is the only occurrence in a schedule, that schedule is reconfigured with no occurrences. If the schedule is included in a maintenance policy and that policy is assigned to a service profile, any pending activities related to the server associated with the service profile cannot be deployed. You must add a one time ocurrence or a recurring occurrence to the schedule to deploy the pending activity.
If this is the only occurrence in a schedule, that schedule is reconfigured with no occurrences. If the schedule is included in a maintenance policy and that policy is assigned to a service profile, any pending activities related to the server associated with the service profile cannot be deployed. You must add a one time ocurrence or a recurring occurrence to the schedule to deploy the pending activity.
If this schedule is included in a maintenance policy, the policy is reconfigured with no schedule. If that policy is assigned to a service profile, any pending activities related to the server associated with the service profile cannot be deployed. You must add a schedule to the maintenance policy to deploy the pending activity.
Procedure
Command or Action
Purpose
Step 1
UCS-A# scope system
Enters system mode.
Step 2
UCS-A /system # delete schedulersched-name
Deletes a scheduler and enters scheduler mode.
Step 3
UCS-A /system # commit-buffer
Commits the transaction to the system configuration.
The following example deletes a scheduler called maintenancesched and commits the transaction:
If you plan to configure this maintenance policy for deferred deployment, create a schedule.
Procedure
Command or Action
Purpose
Step 1
UCS-A#
scope orgorg-name
Enters organization mode for the specified organization. To enter
the root organization mode, type
/ as the
org-name.
Step 2
UCS-A /org #
create maint-policypolicy-name
Creates the specified maintenance policy and enters maintenance policy mode.
Step 3
UCS-A /org/maint-policy #
set reboot-policy{immediate |
timer-automatic |
user-ack}
When a service profile is associated with a server, the server needs to be rebooted to complete the association. Specifying the reboot-policy command determines when the reboot occurs for all service profiles that include this maintenance policy. Possible values include:
immediate--The server reboots as soon as the change is made to the service profile.
timer-automatic--You select the schedule that specifies when maintenance operations can be applied to the server using the set scheduler command. Cisco UCS reboots the server and completes the service profile changes at the scheduled time.
user-ack--The user must explicitly acknowledge the changes by using the apply pending-changes command before changes are applied.
Step 4
UCS-A /org/maint-policy #
set schedulerscheduler-name
(Optional)
If the reboot-policy property is set to timer-automatic, you must select the schedule that specifies when maintenance operations can be applied to the server. Cisco UCS reboots the server and completes the service profile changes at the scheduled time.
Step 5
UCS-A /org/maint-policy #
commit-buffer
Commits the transaction to the system configuration.
The following example creates a maintenance policy called maintenance, sets the system to reboot immediately when a service profile is associated with a server, and commits the transaction:
Deploying a Service Profile Change Waiting for User Acknowledgement
Cisco UCS Manager CLI
cannot deploy all pending service profile changes (for multiple service profiles) waiting for user acknowledgement. To simultaneously deploy all pending service profile changes for multiple service profiles, use Cisco UCS Manager GUI.
Important:
You cannot stop Cisco UCS Manager from rebooting the affected server after you acknowledge a pending activity.
Procedure
Command or Action
Purpose
Step 1
UCS-A# scope orgorg-name
Enters organization mode.
To enter the root organization mode, type / as the org-name.
Step 2
UCS-A /org # scope service-profileprofile-name
Enters organization service profile mode for the specified service.
Deploying a Scheduled Service Profile Change Immediately
Cisco UCS Manager CLI
cannot deploy all scheduled service profile changes (for multiple service profiles) at the same time. To simultaneously deploy all scheduled service profile changes for multiple service profiles, use Cisco UCS Manager GUI.
Important:
You cannot stop Cisco UCS Manager from rebooting the affected server after you acknowledge a pending activity.
Procedure
Command or Action
Purpose
Step 1
UCS-A# scope orgorg-name
Enters organization mode.
To enter the root organization mode, type / as the org-name.
Step 2
UCS-A /org # scope service-profileprofile-name
Enters organization service profile mode for the specified service.