Description

Article explains the scenario where EX4600 device gets stuck in sh-4.1#mode prompt.

Symptoms

  • In case of virtual chassis, show virtual chassis status:

<<<Not Present>>>The status NotPrsnt indicates that the switch is physically disconnected from the existing Virtual Chassis. These devices cannot be managed as a single logical device since they do not have an established physical connection with an existing Virtual Chassis. [EX/QFX] Resolution Guide - Troubleshoot Virtual Chassis (VC)

  • On the console session you can see that the device is in:
-sh-4.1#
  • The below is showing the ending of the boot sequence logs:

/boot: 671502336 bytes were trimmed
Formatting '/junos/images/0/vjunos.img', fmt=qcow2 size=1069318656 backing_file='/recovery/junos/jinstall-qfx-5-flex-13.2X51-D35.3-domestic-201710270626.img' encryption=off cluster_size=65536
No kdump initial ramdisk found.[WARNING]
Rebuilding /boot/boot/initrd-2.6.32-279.22.1.el6.x86_64kdump.img
Starting kdump:[  OK  ]
Starting sshd: [  OK  ]
Starting system logger: [  OK  ]
Starting xinetd: [  OK  ]
Starting sntpc: [  OK  ]
Starting libvirtd daemon: [  OK  ]
Starting vehostd: [  OK  ]
Starting crond: [  OK  ]
Starting watchdog: [  OK  ]
Entering non-interactive startup

localhost login:
localhost login:
login: root
Password:
--- Host 13.2I20150410_0533_dc-builder built 2015-04-10 05:33:47 UTC
-sh-4.1#

13.2I20150410_0533_dc-builder


The above behavior may also occur in Stand Alone after a power failure. This results in the EX4600 switch not being reachable from the management.
When accessing the device via console, it shows the "-sh-4.1" prompt and the Junos CLI commands do not work

-sh-4.1#  show chassis hard
-sh: show: command not found
-sh-4.1# cli
-sh: cli: command not found


The following messages are part of the /var/log/messages.
-sh-4.1# cd /var/log
-sh-4.1# tail -100 messages
localhost vehostd[3696]: vjunos_create: Domain creation failed for vjunos0
localhost vehostd[3696]: vjunos_pending: [vjunos0] creation failed vm state PENDING stay in PENDING
localhost vehostd[3696]: vjunos_vmm_periodic: [vjunos0] vehostd missing keepalive
localhost vehostd[3696]: vjunos_create: MAC for UUID is [d007ca47-d3c0-0500]
localhost kernel: ehci_hcd 0000:00:1d.0: remove, state 1
localhost kernel: usb usb1: USB disconnect, device number 1
localhost kernel: usb 1-1: USB disconnect, device number 2
localhost kernel: ehci_hcd 0000:00:1d.0: USB bus 1 deregistered
localhost kernel: pci-stub 0000:00:1d.0: claimed by stub
localhost kernel: device vnet0 entered promiscuous mode
localhost kernel: eth0br: port 2(vnet0) entering learning state
localhost kernel: device vnet1 entered promiscuous mode
localhost kernel: virbr0: topology change detected, propagating
localhost kernel: virbr0: port 2(vnet1) entering forwarding state
localhost kernel: device vnet2 entered promiscuous mode
localhost kernel: ctrlbr0: port 1(vnet2) entering learning state
localhost kernel: device vnet3 entered promiscuous mode
localhost kernel: eth1br: port 2(vnet3) entering learning state
localhost qemu-kvm: JNPR-PCI-EMU: PCI emulation enabled = 1
localhost qemu-kvm: JNPR-PCI-EMU: Map BCM DMA memory

 

Solution

  • Run the ‘reboot’ command from bash.
  • If the device does not load JunOS, power cycle the device and check.
  • Recover the box with the USB recovery media matching the release. After a format install the device was able to re-integrate on the Virtual-chassis as Prsnt state: [EX/QFX] Procedure to format install QFX5K device using a USB
  • If the device does not power on after the power cycle, proceed with the RMA.

Modification History

2024-05-25 : Article Created
2024-08-21: Stand alone information added

Related Information

On a 5100-VC, Device goes to -sh-4.1# prompt on loading 13.2X51-D35.3 from 16.1R6.1 image
[EX/QFX] Procedure to format install QFX5K device using a USB