Recently I had to upgrade 4500x vss pair to a newer IOS-XE release. While trying to start the ISSU process I came across a few errors. Eventually after fixing those errors I finally managed to start the ISSU process.
Below I recorded some of the headache points i came across and how I fixed them, and also the ISSU process itself.
I started with doing the usual checks on the health status of the vss pair. (redundancy, issu state, license and software state)
Then i tried to start the issu process with:
issu loadversion 1 bootflash:<newimage> 11 slavebootflash:<newimage>
And instantly A nice error came up:
% Active Boot variable is invalid
If this happens make sure to set the current BOOTVAR to the current running IOS version, and the config-reg to value 0x2102!!
Also do not forget to do a write after chaning 1 of the bootvar settings. (after deleting, do a wr.. after adding do a wr.. etc)
In my case the BOOTVAR was set correctly but the error persisted. So I decided to do a force-failover between the active -standby and back again so both units get a reload
before trying the ISSU again.
Now after both units were reloaded and the vss pair was back in its original state i tried the ISSU again.
The process started, the standby node finally rebooted with the NEW ios, and after a while gave an error on the console:
%C4K_REDUNDANCY-2-IOS_VERSION_CHECK_FAIL: IOS version mismatch. Active supervisor version is 15.2(2)E5 (cat4500e-UNIVERSALK9-M). Standby supervisor version is 15.2(2)E8 (cat4500e-UNIVERSALK9-M). Redundancy feature may not work as expected. %C4K_REDUNDANCY-2-NON_SYMMETRICAL_REDUNDANT_SYSTEM: STANDBY:STANDBY supervisor will operate in fallback redundancy mode rpr. %C4K_REDUNDANCY-3-COMMUNICATION: STANDBY:Communication with the peer Supervisor has been established %C4K_REDUNDANCY-2-VS_REBOOT_ON_RPR_FALLBACK: STANDBY:Supervisor in virtual-switch configuration cannot operate in redundancy mode RPR, will be reset %RF-5-RF_RELOAD: STANDBY:Self Reload. Reason: Virtual-switch fallback to RPR %SYS-5-RELOAD: STANDBY:Reload requested by Platform redundancy manager. Reload Reason: Virtual-switch fallback to RPR. Message from sysmgr: Reason Code: Reset Reason:Reset/Reload requested by [console]. [Reload command]
The standby node after this just rebooted back to the old IOS.
This one gave me some headache at first, until someone pointed out to me that I should the ISSU Compatibility matrix here:
For some IOS-XE versions, especially older versions, it is not possible to instantly upgrade to the latest version in a software train/different software train.
After checking the compatibility matrix it showed i could not upgrade from 3.6.5 to 3.6.8 directly but hat to go to 3.6.7 first.
Uploading this image, and restarting the ISSU process, I finally succeeded with using ISSU to update the IOS version..
Below are the normal ISSU steps normally used for Cisco 4500-x switches
ISSU Upgrade steps:
There are two ways to perform an ISSU upgrade: manually, with four commands; or automatically, with one command.
To have more control, the four command method is preferred.
Check the following :
The active and the standby nodes must have the same (supervisor engine) hardware and software versions(Rommon, license features )
Make sure to double check the required rommon version or if a rommon update is needed!
Make a backup of the config and the current IOS image 🙂
The new and old Cisco IOS software images must be loaded into the file systems (bootflash of the active and slavebootflash of the standby)
of both the active and the standby supervisor engines before you begin the ISSU process
bootflash: <– active node
slavebootflash: <– standby node
Check if the redundancy settings are ok.
Standby node should be in STANDBY-HOT mode.
show redundancy swpg001#show redundancy Redundant System Information : ------------------------------ Available system uptime = 48 weeks, 1 hour, 48 minutes Switchovers system experienced = 2 Standby failures = 4 Last switchover reason = user_forced Hardware Mode = Duplex Configured Redundancy Mode = Stateful Switchover Operating Redundancy Mode = Stateful Switchover Maintenance Mode = Disabled Communications = Up Current Processor Information : ------------------------------ Active Location = slot 1/1 Current Software state = ACTIVE Uptime in current state = 20 hours, 28 minutes Image Version = Cisco IOS Software, IOS-XE Software, Catalyst 4500 L3 Switch Software (cat4500e-UNIVERSALK9-M), Version 03.06.05.E RELEASE SOFTWARE (fc2) Technical Support: http://www.cisco.com/techsupport Copyright (c) 1986-2016 by Cisco Systems, Inc. Compiled Thu 02-Jun-16 04:02 by prod BOOT = bootflash:/cat4500e-universalk9.SPA.03.06.05.E.152-2.E5.bin,1; Configuration register = 0x2102 Peer Processor Information : ------------------------------ Standby Location = slot 2/1 Current Software state = STANDBY HOT Uptime in current state = 16 hours, 41 minutes Image Version = Cisco IOS Software, IOS-XE Software, Catalyst 4500 L3 Switch Software (cat4500e-UNIVERSALK9-M), Version 03.06.05.E RELEASE SOFTWARE (fc2) Technical Support: http://www.cisco.com/techsupport Copyright (c) 1986-2016 by Cisco Systems, Inc. Compiled Thu 02-Jun-16 04:02 by pr BOOT = bootflash:/cat4500e-universalk9.SPA.03.06.05.E.152-2.E5.bin,1; Configuration register = 0x2102
Verify ISSU state and write down the slot ID’s of the active and standby node.
In this case the primary node has slot 1 and the secondary node slot 2.
swpg001# show issu state detail Slot = 1 RP State = Active ISSU State = Init Operating Mode = Stateful Switchover Current Image = bootflash:/cat4500e-universalk9.SPA.03.06.05.E.152-2.E5.bin Pre-ISSU (Original) Image = N/A Post-ISSU (Targeted) Image = N/ASlot = 11 RP State = Standby ISSU State = Init Operating Mode = Stateful Switchover Current Image = bootflash:/cat4500e-universalk9.SPA.03.06.05.E.152-2.E5.bin Pre-ISSU (Original) Image = N/A Post-ISSU (Targeted) Image = N/A
Check and if needed set the BOOTVAR to the current running image.
The current running version should be set in the current BOOTVAR.
Also the config-register should be set to 0x2102
Do not forget to write the config if the BOOTVAR is changed!
swpg001#sh bootvar BOOT variable = bootflash:/cat4500e-universalk9.SPA.03.06.05.E.152-2.E5.bin,1; CONFIG_FILE variable does not exist BOOTLDR variable does not exist Configuration register is 0x2102 Standby BOOT variable = bootflash:/cat4500e-universalk9.SPA.03.06.05.E.152-2.E5.bin,1; Standby CONFIG_FILE variable does not exist Standby BOOTLDR variable does not exist Standby Configuration register is 0x2102
*** EXTRA INFO: ***
It is possible that even tho the BOOTVAR is set to the right variable, the ISSU process fails with the error:
“Active BOOT variable is invalid”
In this case, after setting the BOOTVAR again and doing an WR MEM, a forced-failover should be applied 2x
so both nodes are reloaded before attempting ISSU again.
First load the new IOS version on the current STANDBY (slot id 11 in this case) node
issu loadversion 1 bootflash:new_image 11 slavebootflash:new_image
Note: with the first command the STANDBY node will reboot with the NEW ios image.
It may take several seconds after entering the issu loadversion command for Cisco IOS software to load onto the standby supervisor engine and the standby supervisor engine to transition to SSO mode.
If you enter the show issu state command too quickly, you may not see the information you need.
After the standby node is reloaded, we can switch over to this node as the active node:
issu runversion 11 slavebootflash:new_image This command will reload the Active unit. Proceed ? [confirm]
A switchover occurs at this point and the standby node becomes primary active, while the active node reboots with the new image and becomes standby node
if everything seems OK, the rollback timer needs to be stopped.
If this is not done the system automatically aborts the ISSU process and reverts to the original Cisco IOS software version. By default the rollback timer is 45 minutes.
To stop the timer and accept the new running version continue to the next step.
After the previous command and the switchover from the active to the standby node, the standby node became the new active node.
Do not forget that the SLOT ID’s are fixed and move with the switchover. So the new situation is Active node now has SLOT ID 11, and the Standby node SLOT ID 1.
We want to accept the new image running on the now Active node (slot id 11 in this case)
issu acceptversion 11 bootflash:new_image
This will stop the rollback timer and accepts the new image as the new default image.
Now for the last step we need to load the standby supervisor with the new image
issu commitversion 1 slavebootflash:new_image
This will reboot the standby node with the new image.
After the switch is back ‘online’.
Check the following if all states are OK
show redundancy states show redundancy show issu state detail
OPTIONAL: Last step
Do a redundancy force-failover so that the original primary node which now is standby is the active node again.