Power8 VM does not get enlisted in MaaS

Bug #1409952 reported by Ashley Lai
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
MAAS
Fix Released
Critical
Jason Hobbs
1.7
Fix Released
Undecided
Unassigned

Bug Description

The VM was successfully created in libvirt but failed to enlist to MaaS.

https://pastebin.canonical.com/123233/

Related branches

Revision history for this message
Jason Hobbs (jason-hobbs) wrote :

I logged into this node to debug its enlistment - maas-enlist is failing with this message:

ubuntu@maas-enlisting-node:~$ maas-enlist --serverurl http://10.245.0.10/MAAS/api/1.0/nodes/ --hostname 10-245-48-191.oil
{"architecture": ["'ppc64el/chrp_ibm' is not a valid architecture. It should be one of: 'amd64/generic', 'amd64/hwe-p', 'amd64/hwe-q', 'amd64/hwe-r', 'amd64/hwe-s', 'amd64/hwe-t', 'amd64/hwe-u', 'i386/generic', 'i386/hwe-p', 'i386/hwe-q', 'i386/hwe-r', 'i386/hwe-s', 'i386/hwe-t', 'i386/hwe-u', 'ppc64el/generic', 'ppc64el/hwe-p', 'ppc64el/hwe-q', 'ppc64el/hwe-r', 'ppc64el/hwe-s', 'ppc64el/hwe-t', 'ppc64el/hwe-u'."]}

That's because archdetect shows:

ubuntu@maas-enlisting-node:~$ archdetect
ppc64el/chrp_ibm

Should this really be detected as ppc64el/chrp_ibm? Should MAAS treat that subarch just like ppc64el/generic?

tags: added: oil
Revision history for this message
Blake Rouse (blake-rouse) wrote :

I think its detecting the subarch incorrectly or is that a valid subarch for power8?

Changed in maas:
status: New → Incomplete
Revision history for this message
Jason Hobbs (jason-hobbs) wrote :

Turns out I was running the wrong maas-enlist script - there is one from the maas-enlist package too for some reason, which is old and doesn't patch up all ppc64el subarchs to be enlisted as ppc64el/generic, like the new one does.

The new one still failed though, with a different failure:

http://paste.ubuntu.com/9752415/

Not a very heplful error message, but I added some debug prints to the script and figured out that the MAC address list was being submitted as: "52,52:54:00:66:ce:c8"

This is because of code added in https://code.launchpad.net/~andreserl/maas/fix_lp1399676. It incorrectly parses the PXE MAC as '52' because it only matches dashes '-' in MAC addresses, not semicolons.

Changed in maas:
status: Incomplete → Triaged
importance: Undecided → Critical
assignee: nobody → Jason Hobbs (jason-hobbs)
status: Triaged → In Progress
Changed in maas:
milestone: none → 1.7.1
milestone: 1.7.1 → none
Revision history for this message
Christian Reis (kiko) wrote :

That's really interesting; it means that the BOOTIF specified in the kernel commandline (obtained via /proc/cmdline) uses dashes and colons in different situations (and if it's something the kernel uses, I assume the kernel handles both correctly?) What generates the BOOTIF argument?

Revision history for this message
Andres Rodriguez (andreserl) wrote :

This bug has been reported and fixed on upstream MAAS. However, provided that the bug was listed on the debian changelog, this appears as needing verification for pending SRU [1]. This bug did not affect current MAAS in Ubuntu, hence setting this to verification-done to unblock pending SRU.

[1]:http://people.canonical.com/~ubuntu-archive/pending-sru.html

tags: added: verification-done
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.