Technology and Support - Cisco Community



Cisco UCS Firmware Upgrade Procedure – Manual StepsXSJ from 2.2 (1c) -> 2.2 (3f)XIR from 2.2 (2c) -> 2.2 (3f)What's Required for a Firmware Update?There are 2 bundles that are required for our environment: Bundle:ucs-k9-bundle-infra.2.2.3f.A.binB Series Blade Server Bundle: ucs-k9-bundle-b-series.2.2.3f.B.bin??Check List (Check off events as they are completed)Item No.TaskTask StartedTask Completed1Obtaining Software Bundles from Cisco??2Downloading Firmware Images to the Fabric Interconnect from a Local Disk??3Determine the Contents of the Firmware Package??4Review Release Notes??5Back Up the Configuration into an All Configuration Backup File??6Disable Call Home??7Verify and Enable the User Acknowledgement Option in Maintenance Plan for all Service Profiles and Templates if not already done.??8Verify the High Availability Status and Roles of the Cluster Configuration??9Verify the Status of the IO Modules??10Verify the State of Servers??11Verify the Status of Adapters on Servers in the Chassis??12Verify There are No Major Faults in UCS??13Verify the Number of Ethernet Data Paths??14Verify the Data Path for Fibre Channel End-Host Mode??15Enumerate the Current Running/Startup and Backup Configs??16Activate the Cisco UCS Manager Software??17Update the Firmware on the IOMs for each Chassis and both IOM’s in each Chassis??18Activate the Firmware on the IOMs for each Chassis and both IOM’s in each Chassis??19Activate the Firmware on the Subordinate Fabric Interconnect??20Force a Fabric Interconnect Failover??21Verify that the Ethernet Data Path has been Restored??22Verify the Data Path for Fibre Channel End-Host Mode Has Been Restored??23Activate the Firmware on the Primary Fabric Interconnect??24Once the FI A upgrade is completed, then promote FI A the primary Fabric Interconnect??25Create a Host Firmware Package??26Activate the Host Firmware Package??27This is where you evacuate guests from each ESXi server and then reboot servers accordingly28Verify Firmware Versions on Components after reboot??29Enable Call Home30Final verification of all: UCS Infrastructure, ESXi, Servers, VM Guests, and Datastores? NOTE: From here on onwards the photo snapshots are examples for your reference onlyOrder to Upgrade Components?The Infrastructure must be upgraded first and then the Servers.? The specific order within each bundle should be:?Infrastructure Bundle:a.??? UCS Managerb.??? IOMs - The new firmware will be flagged as the "Startup Version".? They will be rebooted once the Interconnect Switches are upgraded and rebootedc.??? Fabric Interconnect Switches?B-Series Blade Server Bundle:a.??? CIMC firmwareb.??? BIOS firmwarec.??? Adapter firmwared.??? Board controller firmware?Method of Upgradea.??? Infrastructure Bundle - Direct upgrade (follow the steps in this procedure)b.??? B-Series Server Bundle - Host Firmware Packages (Follow the steps in this procedure)?1.?? Obtaining Software Bundles from CiscoDetermine which of the following software bundles you need to update the Cisco UCS domain: UCS Infrastructure Software Bundle—Required for all Cisco UCS domains. Ours is 2.2(3F)ucs-k9-bundle-infra.2.2.3f.A.binCisco UCS B-Series Blade Server Software Bundle—Required for all Cisco UCS domains that include blade servers.ucs-k9-bundle-b-series.2.2.3f.B.binVerify that the software integrity by running the checksum md5sum command against each bin you downloaded:# md5sum ucs-k9-bundle-infra.2.2.3f.A.bin3036a71476fea1aeb2f6c72453e4de3d ucs-k9-bundle-infra.2.2.3f.A.bin# md5sum ucs-k9-bundle-b-series.2.2.3f.B.bine54c56a1e92365e6de0b341bb8b13d71 ucs-k9-bundle-b-series.2.2.3f.B.bin?Step 1??? In a web browser, navigate to .Step 2??? Under Support, click All Downloads.Step 3??? In the center pane, click "Servers - Unified Computing"Step 4??? If prompted, enter your username and password to log in.Step 5??? Click "Cisco UCS Infrastructure and UCS Manager Software"Step 6??? Click "Unified Computing System (UCS) Infrastructure Software"Step 7??? Select the Version you are interested in from the Navigation Pane.? Ours is 2.2(3F)Step 7??? Click the Release Notes link to download the latest version of the Release Notes.Step 8??? Download:a.??? ?"Infrastructure Software Bundle"b.??? "Software for the B-Series Blades"?Save the bundles to a location on the network accessible to Cisco UCS??2.?? Downloading Firmware Images to the Fabric Interconnect from a Local Disk?Equipment -->? Equipment?Step 1??? In the Work pane, click the Firmware Management tabStep 2??? Click the Installed Firmware tabStep 3??? Click the "Download Task” TABStep 4 Select “Download Firmware” linka.??? Select the "Local File System" radio buttonb.??? Browse to the location of the firmware bundlec.??? Click "OK"?Repeat the process to upload the Infrastructure and the B-Series Blade Server Bundle?centercenter003.?? Determine the Contents of the Firmware PackageEquipment -->? Equipment?Step 1??? In the Work pane, click the Firmware Management tab.Step 2??? On the Packages sub tab, click the + icon next to a package to view its contents.Step 3?? You can also see the version numbers by clicking the "Images" tab501015182880000?4.?? Review Release Notes .?? Back Up the Configuration into an All Configuration Backup FileAdmin --> General --> Backup Configuration?Step 1??? Click the All option in the Filter dropdown boxStep 2??? In the Backup Configuration dialog box, click Create Backup Operation.Step 3??? In the Create Backup Operation dialog box, do the following:a) Complete the following fields:a.??? Admin State field—Click the Enabled radio button to run the backup operation as soon as you click OK.b.??? Type field—Click the All Configuration radio button to create an XML backup file that includes all system and logical configuration information.c.??? Preserve Identities check box—If the Cisco UCS domain includes any identities derived from pools that you need to preserve, check this check box. Identities such as MAC addresses,?? WWNNs, WWPNs, or UUIDS are assigned at runtime. If you do not want these identities to change after you import the backup file, you must check this check box.? If you do not, these identities may be changed after the import and operations such as a PXE boot or a SAN bootd.??? Location of the Backup File: select "Local File System"e.??? Browse to a location for the file and give the file a name?centercenter006.?? Disable Call HomeCall Home has not been configured in our environment yet on the Cisco Vendor side as firewall rules need to be opened up on Xilinx, so at the moment it is configured on UCS, but not enabled yet.center139255500??7.?? Enable the User Acknowledgement Option in Maintenance Plan for all Service Profiles and Templates415290119062500NOTE: This configuration is should already be in place a standard policy. This is just an additional verification as the default Maintenance Policy has been configured with the Reboot Policy set to "User Acknowledgement".? The policy is bound to all Templates and is part of all Service profiles.?8.?? Verify the High Availability Status and Roles of the Cluster ConfigurationEquipment --> Equipment --> Fabric Interconnects?Step 1??? Click the node for one of the fabric interconnects in the cluster.Step 2??? In the Work pane, click the General tab.Step 3??? If the fields in the "High Availability Details" area are not displayed, click the Expand icon to the right of the heading.Step 4??? Verify that the following fields display the following values:a.??? Ready = Yesb.??? State = Upc.??? Leadership = Either Primary or Subordinate depending on role of Interconnectd.??? Cluster Link Status = Full?Step 5?? Repeat for both Fabric Interconnect nodes?If the values are different, create and download a Tech Support file, and contact Cisco Technical Support. Do not proceed with the firmware upgrade. For more information about Tech Support files, see the Cisco UCS Manager B-Series Troubleshooting Guide.?Step 6??? Note the value in the Leadership field to determine whether the fabric interconnect is the primary or subordinate.? You need to know this information to upgrade the firmware on the fabric interconnects.?Step 7?? Make a note of which Fabric Interconnect Switch is the Primary and which is the Secondarycentercenter9.?? Verify the Status of the IO Modules?Equipment --> Equipment --> Chassis?Step 1??? Click on the chassis for which you want to verify the status of the I/O modules.Step 2??? In the Work pane, click the IO Modules tab.Step 3??? For each I/O module, verify that the following columns display the following values:a.??? Overall Status column = ok?If the values are different, create and download a Tech Support file, and contact Cisco Technical Support. Donot proceed with the firmware upgrade. For more information about Tech Support files, see the Cisco UCSManager B-Series Troubleshooting Guide.?Step 4??? Repeat to verify the status of the I/O modules in each chassis.??centercenter?10. Verify the State of Servers?Equipment --> Equipment --> Chassis?If a server is inoperable, you can proceed with the upgrade for other servers in the Cisco UCS domain. However, you cannot upgrade the inoperable server.?Step 1??? In the Navigation Pane Select a ChassisStep 2??? In the Work pane, click the Servers tab to display a list of all servers in all chassis.Step 3??? For each server, verify that the following columns display the following values:a.??? Overall Status = ok, unassociated, or any value that does not indicate a failure.? If the value indicates a failure, such as discovery-failed, the endpoints on that server cannot be upgraded.b.??? Operability = Operable?Step 4??? If you need to verify that a server has been discovered, do the following:a.??? Right-click the server for which you want to verify the discovery status and choose Show Navigator.b.??? In the Status Details area of the General tab, verify that the Discovery State field displays a value of complete.?If the fields in the Status Details area are not displayed, click the Expand icon to the right of the heading.?Step 5?? Repeat for every chassis full of servers??centercenter?11. Verify the Status of Adapters on Servers in the ChassisEquipment --> Equipment --> Chassis --> {Chassis Number } --> Servers?Step 1??? Click the server for which you want to verify the status of the adapters.Step 2??? In the Work pane, click the Inventory tab.Step 3??? In the Inventory tab, click the Adapters subtab.Step 4??? For each adapter, verify that the following columns display the following values:?Overall Status column = OperableOperability column = Operable?If the fields show a different value and the adapter is inoperable, you can proceed with the upgrade for other adapters on the servers in the Cisco UCS domain. However, you cannot upgrade the inoperable adapter.centercenter?12. Verify There are No Major Faults in UCSCheck all faults to ensure that there isn't anything that will prevent an upgrade.? Resolve all Major Faults where possible.? Servers with Major faults may not be upgraded?13. Verify the Number of Ethernet Data Paths??Command or ActionPurposeStep?1UCS-A /fabric-interconnect # connect nxos {a | b} ? Enters NX-OS mode for the fabric interconnect.Step?2UCS-A(nxos)# show int br | grep -v down | wc –l? Returns the number of active Ethernet interfaces.Step?3UCS-A(nxos)# show platform fwm info hw-stm | grep '1.' | wc –l? Returns the total number of MAC addresses.?The following example returns the number of active Ethernet interfaces and MAC addresses for subordinate fabric interconnect B so that you can verify that the Ethernet data path for that fabric interconnect is up and running:?UCS-A /fabric-interconnect # connect nxos b???UCS-A(nxos)# show platform fwm info hw-stm | grep '1.' | wc -l????14. Verify the Data Path for Fibre Channel End-Host Mode?For best results when upgrading a Cisco UCS domain, we recommend that you perform this task before you begin the upgrade and after you activate the subordinate fabric interconnect, and then compare the two results.??Command or ActionPurposeStep?1UCS-B /fabric-interconnect # connect nxos {a | b} ? Enters NX-OS mode for the fabric interconnect.Step?2UCS-B(nxos)# show npv flogi-table? Displays a table of flogi sessions.Step?3UCS-B(nxos)# show npv flogi-table | grep fc | wc -l? Returns the number of servers logged into the fabric interconnect.?The following example displays the flogi-table and number of servers logged into subordinate fabric interconnect B so that you can verify that the Fibre Channel data path for that fabric interconnect in Fibre Channel End-Host mode is up and running: ssh into your vip pcfiucs01?pcfiucs01 # connect nxos b?pcfiucs01-B (nxos)# show npv flogi-table?pcfiucs01-B(nxos)# show npv flogi-table--------------------------------------------------------------------------------------------------------------------SERVER EXTERNALINTERFACE VSAN FCID PORT NAME NODE NAME INTERFACE--------------------------------------------------------------------------------------------------------------------vfc895 1 0x631507 20:00:00:25:31:05:02:06 20:00:00:25:31:01:00:06 fc2/13vfc1006 1 0x631c02 20:00:00:25:31:05:02:01 20:00:00:25:31:01:00:01 fc2/14vfc1014 1 0x631c03 20:00:00:25:31:05:02:02 20:00:00:25:31:01:00:02 fc2/14 ………….Total number of flogi = 55.pcfiucs01-B(nxos)# show npv flogi-table | grep fc | wc –l55Reference: . Enumerate the Current Running/Startup and Backup ConfigsEquipment --> Equipment --> Firmware Management --> Installed Firmware73812680151100??16. Activate the Cisco UCS Manager Software?Equipment --> Equipment --> Firmware Management --> Installed Firmware?.Step 1Click Activate Firmware. ?Step?2?? On the UCS Manager row of the Activate Firmware dialog box, do the following: a.??? From the drop-down list in the Startup Version column, select the version to which you want to update the software.? We are upgrading to 2.2(3F).b.??? Click OK. ?Cisco UCS Manager disconnects all active sessions, logs out all users, and activates the software. When the upgrade is complete, you are prompted to log back in. If you are prompted to re-login immediately after being disconnected, the login will fail. You must wait until the activation of Cisco UCS Manager is completed, which takes a few minutes.?Cisco UCS Manager makes the selected version the startup version and schedules the activation to occur when the fabric interconnects are upgraded.?Reference: . Update the Firmware on the IOMs[This process will need to be done for each chassis’s IOM’s]Equipment --> Equipment --> Chassis --> Chassis Number --> IO Modules?Step 3??? Click the I/O module that you want to update.Step 4??? In the General tab, click Update Firmware.Step 5??? In the Update Firmware dialog box, do the following:a.??? From the Version drop-down list, select the firmware version to which you want to update the endpointb.??? Click OK?Cisco UCS Manager copies the selected firmware package to the backup memory slot, where it remains untilyou explicitly activate it.?Step 6??? (Optional) Monitor the status of the update in the Update Status area.?The update process can take several minutes. Do not activate the firmware until the selected firmware packagedisplays in the Backup Version field in the Firmware area of the General tab.?centercenter?18. Activate the Firmware on the IOMs[This process will need to be done for each chassis’s IOM’s]?Equipment --> Equipment --> Chassis --> Chassis Number --> IO Modules?Step?1?? Select the IO Module node that includes the I/O module for which you want to activate the updated firmware. Step?2?? In the General tab, click Activate Firmware. Step?3?? In the Activate Firmware dialog box, do the following: a.??? Set the startup version and check the Set Startup Version Only check box. If you configure Set Startup Version Only, the activated firmware moves into the pending-next-reboot state and the endpoint is not immediately rebooted. The activated firmware does not become the running version of firmware until the endpoint is rebooted.b.??? Click OK.c.??? Repeat for all IOMs on all Chassis?********** The IOMs get booted when the Subordinate Fabric Interconnect boots *************?Reference: . Activate the Firmware on the Subordinate Fabric Interconnect?Equipment --> Equipment --> Firmware Management --> Installed Firmware?Step?1?? Determine which fabric interconnect in the cluster is the subordinate fabric interconnect.Step?2?? On the Installed Firmware tab, click Activate Firmware. Step?3?? From the Filter drop-down list on the menu bar, choose Fabric Interconnects.Step?4?? On the row of the Activate Firmware dialog box for the subordinate fabric interconnect, do the following: a.??? In the Kernel row, choose the firmware version to which you want to upgrade from the drop-down list in the Startup Version column.b.??? In the System row, choose the firmware version to which you want to upgrade from the drop-down list in the Startup Version column.Step?5?? Click Apply. Cisco UCS Manager updates and activates the firmware and reboots the fabric interconnect and any I/O module in the data path to that fabric interconnect, disrupting data traffic to and from that fabric interconnect. However, assuming the Cisco UCS domain is configured to permit traffic and port failover, data traffic fails over to the primary fabric interconnect and is not disrupted.Step?6?? Verify the high availability status of the subordinate fabric interconnect. If the High Availability Details area for the fabric interconnect does not show the following values, contact Cisco Technical Support immediately. Do not continue to update the primary fabric interconnect.Field NameRequired ValueReady fieldYes State fieldUp Step?7?? Check the high availability status of the subordinate fabric interconnect.? Ensure it is operable and working correctly?Reference: . Force a Fabric Interconnect FailoverThis operation can only be performed in the Cisco UCS Manager CLI.You must force the failover from the primary fabric interconnect.??Command or ActionPurposeStep?1UCS-A# show cluster state? Displays the state of fabric interconnects in the cluster and whether the cluster is HA ready.? Step?2UCS-A# connect local-mgmt? Enters local management mode for the cluster.? Step?3UCS-A (local-mgmt) # cluster {force primary | lead {a | b}}? Changes the subordinate fabric interconnect to primary using one of the following commands:forceForces local fabric interconnect to become the primary.leadMakes the specified subordinate fabric interconnect the primary.? ?The following example changes fabric interconnect b from subordinate to primary:?UCS-A# show cluster stateCluster Id: 0xfc436fa8b88511e0-0xa370000573cb6c04A: UP, PRIMARYB: UP, SUBORDINATEHA READY?UCS-A# connect local-mgmt?Cisco Nexus Operating System (NX-OS) SoftwareTAC support: (c) 2002-2011, Cisco Systems, Inc. All rights reserved.The copyrights to certain works contained in this software areowned by other third parties and used and distributed underlicense. Certain components of this software are licensed underthe GNU General Public License (GPL) version 2.0 or the GNULesser General Public License (LGPL) Version 2.1. A copy of eachsuch license is available at and(local-mgmt)# cluster lead b?UCS-A(local-mgmt)#Reference: is a screendump of a "Show Cluster State"??21. Verify that the Ethernet Data Path has been restored??Command or ActionPurposeStep?1UCS-B /fabric-interconnect # connect nxos {a | b} ? Enters NX-OS mode for the fabric interconnect.? Step?2UCS-B(nxos)# show int br | grep -v down | wc –l? Returns the number of active Ethernet interfaces.Verify that this number matches the number of Ethernet interfaces that were up prior to the upgrade.? Step?3UCS-B(nxos)# show platform fwm info hw-stm | grep '1.' | wc –l? Returns the total number of MAC addresses.Verify that this number matches the number of MAC addresses prior to the upgrade.? ?The following example returns the number of active Ethernet interfaces and MAC addresses for Primary fabric interconnect B so that you can verify that the Ethernet data path for that fabric interconnect is up and running:?ssh into your vip pcfiucs01?pcfiucs01 # connect nxos b?pcfiucs01-B (nxos)# show platform fwm info hw-stm | grep '1.' | wc -l Reference: . Verify the Data Path for Fibre Channel End-Host Mode Has Been RestoredFor best results when upgrading a Cisco UCS domain, we recommend that you perform this task before you begin the upgrade and after you activate the subordinate fabric interconnect, and then compare the two results.??Command or ActionPurposeStep?1UCS-B /fabric-interconnect # connect nxos {a | b} ? Enters NX-OS mode for the fabric interconnect.? Step?2UCS-B(nxos)# show npv flogi-table? Displays a table of flogi sessions.? Step?3UCS-B(nxos)# show npv flogi-table | grep fc | wc -l? Returns the number of servers logged into the fabric interconnect.The output should match the output you received when you performed this verification prior to beginning the upgrade.? ?The following example displays the flogi-table and number of servers logged into subordinate fabric interconnect B so that you can verify that the Fibre Channel data path for that fabric interconnect in Fibre Channel End-Host mode is up and running: ?ssh into your vip pcfiucs01?pcfiucs01 # connect nxos b?pcfiucs01-B (nxos)# show npv flogi-table?pcfiucs01-B(nxos)# show npv flogi-table--------------------------------------------------------------------------------------------------------------------SERVER EXTERNALINTERFACE VSAN FCID PORT NAME NODE NAME INTERFACE--------------------------------------------------------------------------------------------------------------------vfc895 1 0x631507 20:00:00:25:31:05:02:06 20:00:00:25:31:01:00:06 fc2/13vfc1006 1 0x631c02 20:00:00:25:31:05:02:01 20:00:00:25:31:01:00:01 fc2/14………….Total number of flogi = 55.pcfiucs01-B(nxos)# show npv flogi-table | grep fc | wc –l55Reference: ?23. Activate the Firmware on the Primary Fabric Interconnect?This procedure continues directly from Activating the Firmware on a Subordinate Fabric Interconnect and assumes you are on the Firmware Management tab.?Step?1?? On the Installed Firmware tab, click Activate Firmware. Step?2?? From the Filter drop-down list on the menu bar, choose Fabric Interconnects.Step?3?? On the row of the Activate Firmware dialog box for the subordinate fabric interconnect, do the following: a.??? In the Kernel row, choose the firmware version to which you want to upgrade from the drop-down list in the Startup Version column.b.??? In the System row, choose the firmware version to which you want to upgrade from the drop-down list in the Startup Version column.Step?4?? Click Apply. Cisco UCS Manager updates and activates the firmware and reboots the fabric interconnect and any I/O module in the data path to that fabric interconnect, disrupting data traffic to and from that fabric interconnect. However, assuming the Cisco UCS domain is configured to permit traffic and port failover, data traffic fails over to the other fabric interconnect, which becomes the primary. When it comes back up, this fabric interconnect is the subordinate fabric interconnect.Step?5?? Verify the high availability status of the fabric interconnect. If the High Availability Details area for the fabric interconnect does not show the following values, contact Cisco Technical Support immediately.?Field NameRequired ValueReady fieldYes State fieldUp ?Reference: . Once the FI A upgrade is completed, then promote FI A the primary Fabric Interconnect.Promotion of FI A Steps==================# ssh admin@pcfiucs01bpcfiucs01-B# connect local-mgmt pcfiucs01-B(local-mgmt)# cluster lead aCluster Id: 0x641eff9c10dc11e3-0xaf19002a6a54b564Check (Will show take over in progress, so wait till its s completed!)================================================pcfiucs01-B(local-mgmt)# show cluster stateCluster Id: 0x641eff9c10dc11e3-0xaf19002a6a54b564B: UP, SUBORDINATE, (Management services: SWITCHOVER IN PROGRESS)A: UP, PRIMARY, (Management services: SWITCHOVER IN PROGRESS)HA NOT READYManagement services: switchover in progress on local Fabric InterconnectCheck momentarily later==================pcfiucs01-B(local-mgmt)# show cluster stateCluster Id: 0x641eff9c10dc11e3-0xaf19002a6a54b564B: UP, SUBORDINATEA: UP, PRIMARYHA READYpcfiucs01-B(local-mgmt)# exit 25. Create a Host Firmware PackageServers --> Servers --> Policies --> RootStep?1?? Right-click Host Firmware Packages and choose Create Package. Step?2?? In the Create Host Firmware Package dialog box, enter a unique name and description for the package.? Step?3?? Leave the "Simple" radio button selectedStep?4?? Select the 2.2(3F) version from the "Blade Package" drop down box and click "OK"Step5Select the new Host Firmware Package in the navigation pane and you will see all of the options that have been automatically selected in the content pane for hardware found in the UCS environment?NOTE: For XSJ this step is completed!centercentercentercenter?26. Activate the Host Firmware Package?Servers --> Service Profile Templates --> Root --> Sub-Organizations --> xsjgops or xirgops?Step?1?? Click the service profile template to which you want to add the firmware packages.Step?2?? In the Work pane, click the Policies tab.Step?3?? Click the down arrows to expand the Firmware Policies section.Step?4?? To add a host firmware package, select the desired policy from the Host Firmware drop-down list.Step?5?? Do these steps to each of your service profile template.Step?6?? Click Save Changes.Step?7?? You will have to restart the servers manually using the "Pending Activities" button because we have configured a Maintenance Policy to ask for User Acknowledgement for a reboot eference: ? 27.Critical Step: RebootingAfter firmware package is selected to your “service profile template”, this is where you must evacuate guests from each ESXi server and then reboot servers accordingly. 28. Verify Firmware Versions on Components?Equipment --> Equipment --> Firmware Management --> Installed Firmware?Step?1?? On the Installed Firmware tab, review the firmware versions listed for each component.?Reference: . Enable Call Home?Enable Call home back again. 30. Final verification of all: UCS InfrastructureESXi ServersVM GuestsDatastores > UCS Cluster UPGRADE COMPLETED < ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download