Upgrade 3850 from 3.6.4E to Fuji Bundle and Install mode fail

[sp2720401]

There is a limitation when going from 3.X to 16.3.5 Denali because of the 3.Xe boot loader being unable to boot a file more than 500MB from the flash.

I cannot get 3.6.4E to update to Fuji 16.9.8. That OS file is only 429MB so that limitation should not apply.

Is there an actual update path to get there?

Should I actually update to a later 3.X (3.6.9E) then go to 16.9.8 or should I upgrade to Denali 16.1.1 first then upgrade to Fuji?

[vencislav.metev]

Hi,

You might need to upgrade to a lower version first and then to the version which you want.

Check this article I might help you:

https://www.cisco.com/c/en/us/td/docs/ios-xml/ios/16_xe/Denali-16-3-migration-guide-for-access-and-edge-routers.html#con_999972

Regards,

Ventsi

[vencislav.metev]

Hi,

Or this one which is for 3850:

https://www.cisco.com/c/en/us/td/docs/switches/lan/catalyst3850/software/release/16-1/release_notes/ol-16-1-3850.html#pgfId-1135768

Regards,

Ventsi

[Leo Laohoo]

---

@sp2720401 wrote:

There is a limitation when going from 3.X to 16.3.5 Denali because of the 3.Xe boot loader being unable to boot a file more than 500MB from the flash.

I cannot get 3.6.4E to update to Fuji 16.9.8. That OS file is only 429MB so that limitation should not apply.

---

This is a bug with 3.X.X and affects stacks with very high uptime. 

Was the flash of the stack "cleaned up" before unpacking the package files:  software clean force

Cisco 3850: IOS-XE/Firmware Upgrade

[sp2720401]

The SHA512 checksum performed matches for file cat3k_caa-universalk9.16.09.08.SPA.bin

The flash was cleaned and the switch stack has been rebooted recently due to the following bug for high switch up time. 

Field Notice: FN - 64191 - Catalyst 3850/3650 Switches That Run Cisco IOS XE Releases 3.6.5E/3.7.4E or Earlier Might Exhibit High CPU Usage - Workaround Provided - Cisco

In Install mode I get the following error.

*Jan 24 2022 04:09:08.302 MST: %IOSXE-7-PLATFORM: 2 process xlogger: [packtool]:(debug): Received script exit status 1 [1]: % flash:cat3k_caa-universalk9.16.09.08.SPA.bin does not appear to be a superpackage. Operation aborted.

In bundle mode I get the following error.

File "flash:cat3k_caa-universalk9.16.09.08.SPA.bin" uncompressed and installed, entry point: 0x816e5650
Loading Linux kernel with entry point 0x816e5650 ...
Bootloader: Done loading app on core_mask: 0xf

### Launching Linux Kernel (flags = 0x5)

● mount_packages.service - Mount packages
Loaded: loaded (/lib/systemd/system/mount_packages.service; enabled; vendor preset: enabled)
Active: failed (Result: exit-code) since Mon 2022-01-24 11:22:14 Universal; 121ms ago
Process: 546 ExecStart=/mount_packages.sh (code=exited, status=1/FAILURE)
Main PID: 546 (code=exited, status=1/FAILURE)

[ 12.412361] localhost systemd[1]: Starting Mount packages...
[ 12.461849] localhost mount_packages.sh[546]: booted_item=/flash/cat3k_caa-universalk9.16.09.08.SPA.bin
[ 12.471734] localhost mount_packages.sh[546]: bootloader_version=
[ 13.564129] localhost mount_packages.sh[546]: Releasing iso memory
[ 13.773681] localhost mount_packages.sh[546]: /usr/bin/unbundle
[ 13.778539] localhost mount_packages.sh[546]: nova super packaging
[ 13.782631] localhost mount_packages.sh[546]: rp_super
[ 17.299995] localhost mount_packages.sh[546]: umount: /mnt/piso: not mounted
[ 17.911372] localhost mount_packages.sh[546]: packages.conf header isn't as expected.
[ 17.912105] localhost mount_packages.sh[546]: it must not be a proper packages.conf
[ 17.913117] localhost systemd[1]: mount_packages.service: Main process exited, code=exited, status=1/FAILURE
[ 17.914220] localhost systemd[1]: Failed to start Mount packages.
[ 17.975216] localhost systemd[1]: mount_packages.service: Unit entered failed state.
[ 17.975979] localhost systemd[1]: mount_packages.service: Failed with result 'exit-code'.
Jan 24 11:22:14.904: %BOOT-3-SYSD_STARTFAIL: R0/0: Failed to launch boot task mount_packages.service ( exit-code )
>>> Rebooting

[sp2720401]

I posed a longer response and got a fancy 20 post badge for it. It is gone now for some reason.

the switch stack has been rebooted due to a high CPU field notice for 6.6.4E

The software clean command was issued.

The system uptime is less than 4 days right now it was freshly rebooted before the installation attempt. The install mode failed with an invalid superpackage error. The bundle mode failed with aa invalid packages.conf error.

The SHA512 checksum matches what is provided at the Cisco download portal for the file I am trying to load.

[Leo Laohoo]

---

@sp2720401 wrote:

Active: failed (Result: exit-code) since Mon 2022-01-24 11:22:14 Universal; 121ms ago
Process: 546 ExecStart=/mount_packages.sh (code=exited, status=1/FAILURE)
Main PID: 546 (code=exited, status=1/FAILURE)

---

I have seen these annoying 3 lines so many times.  

I have some good news and I have some bad news.  

Bad News:  I hope the config is backed up

Good news:  It can be fixed. 

1. Force the switch to boot into ROMMON. 
2. Read Cisco 3850: IOS-XE/Firmware Upgrade
3. While the switch is in ROMMON, use Emergency-Install to recover the switch.

[sp2720401]

I have a maintenance scheduled for next week. I will see if the emergency install will work. 

Thanks.

[Leo Laohoo · Accepted solution]

---

@sp2720401 wrote:

I have a maintenance scheduled for next week. I will see if the emergency install will work. 

---

Remember, the Emergency-Install procedure will erase the contents of the flash, including the config and the VLAN.dat file.  
Make sure to have the latest config file ready.

[sp2720401]

Upgraded just fine. emergency-install was the way to go.

[Leo Laohoo]

Thanks for the update.  I am happy to see everything is now working.