Upgrading Clustered Controllers (CLI)
-
Complete Preparing for a Software Upgrade.
Note:
For the purpose of this procedure, the first controller to be upgraded is referred to as controller A and its peer is controller B. If one of the controllers is in a Stripped state (it has no active resources), upgrade that controller first to avoid availability delays. If both controllers in a cluster have active resources, choose either controller to upgrade first.- Log in to controller A, and failover its resources to controller B by performing a reboot:
- Click on the power icon in the upper left section of the navigation bar.
- Select Reboot from the drop-down menu, and click OK to confirm.
Do not select Diagnostic Reboot. Wait for the reboot to complete before proceeding.
- On controller A, go to
maintenance system updates
. - Enter
select
and the name of the update you want to install.controller-a:maintenance system updates> select ak-nas@2013.06.05.6.3,1-2.1.7.1
- Enter
check
, and then entery
to perform health checks.controller-a:maintenance system updates ak-nas@2013.06.05.6.3,1-2.1.7.1> check You have requested to run checks associated with waiting upgrade media. This will execute the same set of checks as will be performed as part of any upgrade attempt to this media, and will highlight conditions that would prevent successful upgrade. No actual upgrade will be attempted, and the checks performed are of static system state and non-invasive. Do you wish to continue? Are you sure? (Y/N) y
- Enter
upgrade
, and then entery
to begin the update process.Note:
You can cancel the update at any time in the CLI by pressing Control-C. If an update problem is marked as repaired during an upgrade from software release OS8.8.0 to a later release, the upgrade completion notice is not sent.As part of the update, controller A automatically reboots.
controller-a:maintenance system updates ak-nas@2013.06.05.6.3,1-2.1.7.1> upgrade This procedure will consume several minutes and requires a system reboot upon successful update, but can be aborted with [Control-C] at any time prior to reboot. A health check will validate system readiness before an update is attempted, and may also be executed independently using the check command. Are you sure? (Y/N) y
- After controller A has fully rebooted, log back in to controller A, and enter
configuration cluster show
to verify that controller A is in theReady (waiting for failback)
state.In the following example, the state of controller A is
AKCS_STRIPPED
, with a description ofReady (waiting for failback)
.controller-a:> configuration cluster show state = AKCS_STRIPPED description = Ready (waiting for failback) ... peer_state = AKCS_OWNER peer_description = Active (takeover completed)
For information about the cluster states, see Cluster Takeover and Failback in Oracle ZFS Storage Appliance Administration Guide, Release OS8.8.x.
- Enter
maintenance system updates firmware show
to monitor firmware updates.controller-a:> maintenance system updates firmware show Pending No Pending Updates Failed No Failed Updates In Progress No Updates in Progress
- Log in to controller B, and enter
configuration cluster show
to verify that controller B is in theActive (takeover completed)
state.For information about the cluster states, see Cluster Takeover and Failback in Oracle ZFS Storage Appliance Administration Guide, Release OS8.8.x.
In the following example, the state of controller B is
AKCS_OWNER
, with a description ofActive (takeover completed)
.controller-b:> configuration cluster show state = AKCS_OWNER description = Active (takeover completed) ... peer_state = AKCS_STRIPPED peer_description = Ready (waiting for failback)
- Enter
configuration cluster failback
, and then entery
to change the cluster to an Active/Active configuration.Note:
This is not necessary if you want an Active/Passive configuration.controller-b:> configuration cluster failback Continuing will immediately fail back the resources assigned to the cluster peer. This may result in clients experiencing a slight delay in service. Are you sure? (Y/N) y
- Enter
maintenance system updates
, and entershow
to view the available updates.controller-b:> maintenance system updates hostname:maintenance system updates> show Updates: UPDATE RELEASE DATE STATUS ak-nas@2013.06.05.6.2,1-1.1 2016-7-26 15:20:06 current ak-nas@2013.06.05.6.3,1-2.1.7.1 2016-8-24 14:10:02 waiting ak-nas@2013.06.05.6.1,1-1.1 2016-6-24 23:37:50 previous
- Enter
select
and the name of the update you want to install.controller-b:maintenance system updates> select ak-nas@2013.06.05.6.3,1-2.1.7.1
- Enter
check
, and then entery
to perform health checks.controller-b:maintenance system updates ak-nas@2013.06.05.6.3,1-2.1.7.1> check You have requested to run checks associated with waiting upgrade media. This will execute the same set of checks as will be performed as part of any upgrade attempt to this media, and will highlight conditions that would prevent successful upgrade. No actual upgrade will be attempted, and the checks performed are of static system state and non-invasive. Do you wish to continue? Are you sure? (Y/N) y
- Enter
upgrade
, and then entery
to begin the update process.Note:
You can cancel the update at any time in the CLI by pressing Control-C. If an update problem is marked as repaired during an upgrade from software release OS8.8.0 to a later release, the upgrade completion notice is not sent.As part of the update, controller B automatically reboots.
controller-b:maintenance system updates ak-nas@2013.06.05.6.3,1-2.1.7.1> upgrade This procedure will consume several minutes and requires a system reboot upon successful update, but can be aborted with [Control-C] at any time prior to reboot. A health check will validate system readiness before an update is attempted, and may also be executed independently using the check command. Are you sure? (Y/N) y
- After controller B has fully rebooted, log back in to controller B, and enter
configuration cluster show
to verify that controller B is in theReady (waiting for failback)
state.For information about the cluster states, see Cluster Takeover and Failback in Oracle ZFS Storage Appliance Administration Guide, Release OS8.8.x.
In the following example, the state of controller B is
AKCS_STRIPPED
, with a description ofReady (waiting for failback)
.controller-b:> configuration cluster show state = AKCS_STRIPPED description = Ready (waiting for failback) ... peer_state = AKCS_OWNER peer_description = Active (takeover completed)
- Enter
maintenance system updates firmware show
to monitor firmware updates.controller-b:> maintenance system updates firmware show Pending No Pending Updates Failed No Failed Updates In Progress No Updates in Progress
- Log in to controller A, and enter
configuration cluster show
to verify that controller A is in theActive (takeover completed)
state.In the following example, the state of controller A is
AKCS_OWNER
, with a description ofActive (takeover completed)
.controller-a:> configuration cluster show state = AKCS_OWNER description = Active (takeover completed) ... peer_state = AKCS_STRIPPED peer_description = Ready (waiting for failback)
For information about the cluster states, see Cluster Takeover and Failback in Oracle ZFS Storage Appliance Administration Guide, Release OS8.8.x.
- To verify that all firmware updates are complete, enter
maintenance system updates firmware show
to monitor firmware updates.controller-a:> maintenance system updates firmware show Pending No Pending Updates Failed No Failed Updates In Progress No Updates in Progress
Note:
Do not begin the next step until all firmware updates are complete. - Enter
configuration cluster failback
, and then entery
to change the cluster to an Active/Active configuration.Note:
This is not necessary if you want an Active/Passive configuration.controller-a:> configuration cluster failback Continuing will immediately fail back the resources assigned to the cluster peer. This may result in clients experiencing a slight delay in service. Are you sure? (Y/N) y
Both controllers are now upgraded.
- Enter
maintenance hardware show
to verify that all disks are online.controller-a:> maintenance hardware show Chassis: NAME STATE MANUFACTURER MODEL SERIAL RPM ... disk-000 HDD 0 ok SEAGATE ST330057SSUN300G 001117G1G929 6SJ1G929 15000 disk-001 HDD 1 ok SEAGATE ST330057SSUN300G 001117G1GA8Y 6SJ1GA8Y 15000 disk-002 HDD 2 ok SEAGATE ST330057SSUN300G 001117G1KDPZ 6SJ1KDPZ 15000 ...
- Enter
configuration storage show
to check the status of each pool.Pools should be online and have no errors.
controller-a:> configuration storage show Pools: POOL OWNER DATA PROFILE LOG PROFILE STATUS ERRORS -> pool25 controller-a - - online 0 pool26 controller-b - - online 0
- Log in to controller B, and repeat steps 18-19 for controller B.
- Enable any data services that were disabled before the upgrade, as described in Enabling a Service (CLI) in Oracle ZFS Storage Appliance Administration Guide, Release OS8.8.x.