Routing Engine may fail to boot online after Junos Downgrade. This is more applicable for Routing Engine RMA cases where customer is running older release and they receive new RMA RE with latest wrk version. In this case, the normal downgrade to customers qualified older release can hit this issue
Kernel Version showing as LTS19
root> show vmhost snapshot
Nov 03 01:42:19
UEFI Version: CBEP_P_VAL1_00.18.02
Secondary Disk, Snapshot Time: <fresh install>
Version: set p
VMHost Version: 7.2534
VMHost Root: vmhost-x86_64-21.4R3-20220823_0241_builder
VMHost Core: vmhost-core-x86-64-21.4R3-S2.3
kernel: 5.2.60-rt15-LTS19
Junos Disk: junos-install-mx-x86-64-21.4R3-S2.3
Version: set b
Below errors are usually reported in Routing Engine Console
…
error: bad signature 0x0f.
error: you need to load the kernel first.
Press any key to continue...
It has been observed that if the disk layout is initially created with LTS19 based FS tools then the subsequent downgrade to WRL6 based releases fail to boot due to FS tools incompatibility which is brought in by LTS19 based tooling versions. The rollback mechanism from initramfs also fails because of the missing signature files.
Basically a normal step downgrade from LTS19 to WRL6 works fine but if the initial disk layout has been created with utilities based of LTS19, then we see this issue.
Option 1: Recover Routing Engine using bootable USB media
Option 2: Please follow below steps for non USB media recovery
- Select Disk 1 (Secondary Disk) as my primary boot medium from Routing Engine BIOS menu.
- Install a release running WRL9 as intermeidate version and reboot the system. Junos Software will get added to Disk 0 (Primary Disk)
- Device will now boot from Disk 0 (Primary Disk)- Junos running on WRL9
- Create VMhost Snapshot to reimage Disk 1 (Secondary Disk) to WRL9 based Junos
- Reboot using “request vmhost reboot disk2”. This is to reboot Routing Engine from Disk 1 (Secondary Disk) running WRL9 based Junos
- Install the customers qualified WRL6 based Junos release and reboot
- Lastly, create “VMhost Snapshot” to reimage Disk 1 (Secondary Disk)