Comment 6 for bug 1859656
Revision history for this message
| Sean Feole (sfeole) wrote Re: [2.6] Unable to reboot s390x KVM machine after initial deploy | #6 |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
2eb648a
demo site
(Get the code!)
Hey Lee,
I took a look at that document. I want to make a few points here. This has worked in the past, earlier versions of MAAS. Nothing has ever changed on my lpar that is hosting the VM's.
The host lpar is Bionic, 18.04. This was working for months and suddenly stopped, the only change in my labs have been me updating the maas servers to newer versions.
According to libvirt documentations the s390x arch only respects the first <OS> <boot> param in the XML.
<os>
<type arch='s390x' machine=
<boot dev='network'/>
<boot dev='hd'/>
</os>
In normal circumstances, net boot fails and the VM default to the HD, but on s390 that's not the case. Once net boot fails then
Connected to domain s2lp6g001
Escape character is ^]
done
Using IPv4 address: 10.246.75.177
Using TFTP server: 10.246.72.3
Bootfile name: 'boots390x.bin'
Receiving data: 0 KBytes
TFTP error: file not found: boots390x.bin
Trying pxelinux.cfg files...
Receiving data: 0 KBytes
Receiving data: 0 KBytes
Failed to load OS from network
your stuck there in the off state forever. Changing the XML so that the VM boots from <HD> first works. however that's not really acceptable in this use case.
I would imagine that for this to work MAAS-dhcp would have to instruct the s390x VM to boot from "local" (disk) once it's already deployed. All by means of the /var/lib/
Is that not how this is designed to work?