[solved] Boot loop with 18.04


#1

Hi, I manage a cluster using MAAS. I’m starting to upgrade the nodes from 14.04 to 18.04, and some nodes are getting stuck in a boot loop with 18.04. Specifically, when the node gets to:

Loading ubuntu/amd64/generic/bionic/daily/boot-initrd… ok

The node reboots immediately. This repeats once it reboots. Here are some key details:

  1. This occurs during Commissioning using 18.04. I can get Commissioning to work by switching to 16.04. However, I get the same problem when I go to Deploy 18.04.

  2. The issue occurs every time with a subset of nodes. However, other nodes with identical hardware do not have this problem.

Could this be related to the fact that I had a messy upgrade of my MAAS server from 14.04 to 16.04. Is there a good way to install a new MAAS server from scratch and point the nodes to it?

Sorry for a newbie question, but I’ve lost several days trying to debug this issue on my own.


#2

Problem solved: when I commissioned some nodes, I decided to ‘remove’ the second NIC from the MAAS list of interfaces. That really screwed up the underlying hardware - the underlying Supermicro BMC reported it only had 1 physical NIC. For some reason, 16.04 (Xenial) did not complain, but when I booted 18.04 (Bionic), it got into a reboot loop each time it loaded the initial ramdisk (boot-initrd).

Seems like there is an upstream problem with Bionic, though it seems like an edge case.


#3

I have same problem and I don’t understand exact reason for this, I didn’t do anything with interfaces in node configuration, but it still stuck in boot loop in ubuntu 18.04 (and works fine in 16.04). Can you add some details for this?