Recovering an Instance/Virtual Machine (VM) that is refusing to boot
- Navigate to the Nirin Dashboard at cloud.nci.org.au
- Click on the Instance which does not boot.
- Click on the Console tab and open the console in a separate browser window.
- In the Console window click on the Send Ctrl-Alt-Del button
- Wait for the Instance to reboot to the GRUB menu
- Press 'e' when one of the GRUB boot option lines are selected
- Scroll down to the Kernel Command Line
- Remove the text console=ttyS0,115200 to ensure that the (novnc based) Console is able to accept IO from keyboard
- Add a relevant option to drop down to a recovery terminal
- If you want to drop down to your ramdisk image with no filesystems mounted, add rd.break to the kernel command line.
- If you want a bash prompt with the rootfs mounted, add init=/bin/bash
- Press Ctrl-x to boot with modified kernel command line
- You should see a bash prompt after some initial boot steps. At this point if you need to make changes to any files on the rootfs, please remount it with the rw flag:
mount -o remount,rw /
- Make the changes as required. For example, if you have an issue with a /g/data mount included in /etc/fstab, open /etc/fstab and comment the problematic line
touch ~/.autorelabel
exec /sbin/init
- Your Instance should now boot