You can upgrade the firmware on one or more compute devices in a system to the latest version. For both switches and servers you can upgrade the firmware from the prior two versions.
For HA800 G2/G3 series servers, the I/O cards are automatically discovered when they are onboarded either as bare metal servers or with ESXi details.
For a table of supported compute BMC firmware upgrade paths from the previous two versions, see Supported firmware upgrade paths for compute devices
- Upload the firmware bundle to UCP Advisor. For details, see Uploading firmware. For Brocade switch firmware, see Uploading firmware for Brocade Fibre Channel G620 and G720 switches.
- For Ethernet switches, verify that the device user credentials are assigned administrator role: network-admin.
- For Fibre Channel switches, verify that the device user credentials are assigned administrator role: admin
- For compute nodes:
- Verify that the device user credentials are assigned administrator role: Administrator.
- Browse to the UCP system compute inventory and then refresh it. Reset the BMC for any compute node that does not show Normal status. See Resetting the BMC of a single compute node.
- Place the compute nodes in maintenance mode before performing parallel upgrades.
- For I/O cards:
- For DS120/DS220/DS225/DS240 and DS120 G2/DS220 G2 server models, you must register the vCenter and attach it to the UCP system that has the compute servers onboarded with ESXi details.
- For HA800 G2/G3 series servers, the I/O card upgrade works for both bare metal servers and servers with ESXi details.
- Verify that a vCenter is registered and attached to the UCP system.
- Verify the currently supported I/O cards. See: Supported I/O cards: DS120/DS220/DS225/DS240 and DS120/DS220 G2 server models.
- Add the ESXi host to the vCenter where the UCP system is registered so that the I/O cards can be discovered.
- Place the compute nodes in maintenance mode before performing parallel I/O card discovery.
- Onboard the servers to UCP Advisor and then verify that the BMC and ESXi details are available.
- (Optional for servers only) Turn on the Upgrade in parallel toggle to upgrade the selected non-management servers in parallel. You can upgrade up to ten servers at a time in parallel. Any servers over ten are processed in a subsequent batch.Note:
- For HA800 G2/G3 series servers, the following upgrade options are supported:
- Upgrade all firmware: Select this option to upgrade all the server components, such as iLO, BIOS, and I/O cards. This is the default option.
- Upgrade BIOS and iLO firmware only: Select this option to upgrade only the iLO and the BIOS components.
- Upgrading the firmware on multiple compute devices in parallel reduces the downtime of the devices during the upgrade.
- By default, the servers are upgraded sequentially. In this method, all the selected compute devices are upgraded sequentially in the order that they are listed.
- Parallel or sequential firmware upgrade is not supported on Ethernet switches and Fibre Channel switches.
- Before upgrading the compute firmware or discovering the I/O cards, the BMC is reset first, which will take at least 7 minutes to complete.
Important: Running the firmware upgrade operation for more than ten servers at a time may cause the operation for some servers to time out before completing. - For HA800 G2/G3 series servers, the following upgrade options are supported:
For a sequential upgrade, if the cause of the failure is specific to the server where the task terminated, reinitiate the task for the remaining servers. You can manually update the firmware on the failed server using the server's BMC UI.
During a sequential upgrade or a parallel upgrade, an individual task is created for each device on the Tasks tab. The Initiate Upgrade server firmware task displays the details of all the servers selected for the upgrade.
Monitor the Tasks tab to see which cards are supported and which cards are being upgraded.
After the I/O card upgrade completes, the card inventory is updated in the compute I/O card inventory tab.