Template boot order is not converted during upgrade to 5.0
|Assignee:||Carlos Martín||% Done:|
|Category:||Core & System|
|Target version:||Release 5.2|
|Affected Versions:||OpenNebula 5.0|
Hello, follow up of my forum post.
I just upgrade our production 4.14 OpenNebula to 5.0.2 following the documentation.
I just have an issue with already existing templates:
Thu Aug 18 16:33:30 2016 [Z0][ReM][E]: Req:8464 UID:5 TemplateInstantiate result FAILURE [TemplateInstantiate] Error allocating a new virtual machine template. Wrong OS/BOOT value: "hd" should be a comma-separated list of disk# or nic#
I thought this would be managed by the
I made a
onedb patch file to fix most of the cases, I only skip templates when the image referenced by a
DISK is not found, so I can't check the index
Here is the log of the patch command:
- Skipping some template because of missing image:
Skipping template 1380 Install CDROM 40Go (i386): No image named 'eole-2.4.1-rc2-i386.iso' for uid '9' Skipping template 1381 Install CDROM 40Go (amd64): No image named 'eole-2.4.1-rc2-amd64.iso' for uid '9'
- Updating a template after skipping an nonexistant boot device type:
Warning template 13219 coreos-amd64: skip nonexistent boot device 'cdrom' Template 13219 coreos-amd64: 'hd,cdrom' -> 'disk0'
- Updating a template without any problem, note that the first boot device
CDROMis the second
Template 2349 Run Any: 'CDROM,hd' -> 'disk1,disk0'
Skipping messages are not conditionned on
ops[:verbose], it may even requires a dedicated logger for admin to follow what's happening.