Bug 1324447

Moving from 4.0 alpha to 4.0 beta since 4.0 alpha has been already released and bug is not ON_QA.

Looking in the XMLRPC struct returned to the engine I see that we look for a "operatingSystem" field in that structure 
While debugging the code I go the following values for "operatingSystem":

release" -> "2.1511.el7.centos.2.10
version" -> "7
name" -> "RHEL

On he other hand , looking in the os_release file on the node I see the following :

NAME="CentOS Linux"
VERSION="7 (Core)"
ID="centos"
ID_LIKE="rhel fedora"
VERSION_ID="7"
VARIANT="oVirt Node 4.0.0_master"
VARIANT_ID="ovirt-node"
PRETTY_NAME="oVirt Node 4.0.0_master"
ANSI_COLOR="0;31"
CPE_NAME="cpe:/o:centos:centos:7"
HOME_URL="https://www.ovirt.org/"
BUG_REPORT_URL="https://bugzilla.redhat.com/"

So, it seems that the problem origin is in VDSM

oVirt 4.0 beta has been released, moving to RC milestone.

oVirt 4.0 beta has been released, moving to RC milestone.

The propper - and upstream aligned - handling is probably to include the id and variant_id fields of the os-release file.

This tuple should allow to identify hosts reliably.

*** Bug 1332466 has been marked as a duplicate of this bug. ***

We decided to add pretty-name to current host OS identification (name, version, release) reported by VDSM ti distinguish between EL7 and NGN based on EL7.

Moving back to POST as backport to ovirt-engine-4.0 branch is needed

All patches are merged and there should be no risk around stability in those patches, so moving to 4.0.2

Setting target release to 4.18.8 as it's contained in latest VDSM build for ovirt-4.0.2-RC1

vdsm returns pretty name

	operatingSystem = {'name': 'RHEL',
	                   'pretty_name': 'Red Hat Virtualization Host 4.0 (el7.2)',
	                   'release': '1.el7',
	                   'version': '4.0'}


verified in vdsm-4.18.11-1.el7ev.x86_64

this is actually wrong, it should be release=$RHEL_RELEASE, version=$RHEL_VERSION

Can you paste /etc/os-release, /etc/system-release, /etc/system-release-cpe and /etc/redhat-release  of the host?

Only pretty name should contain the RHV version/release

Target release should be placed once a package build is known to fix a issue. Since this bug is not modified, the target version has been reset. Please use target milestone to plan a fix for a oVirt release.

# cat /etc/os-release
NAME="Red Hat Enterprise Linux"
VERSION="7.2"
VERSION_ID="7.2"
ID="rhel"
ID_LIKE="fedora"
VARIANT="Red Hat Virtualization Host"
VARIANT_ID="ovirt-node"
PRETTY_NAME="Red Hat Virtualization Host 4.0 (el7.2)"
ANSI_COLOR="0;31"
CPE_NAME="cpe:/o:redhat:enterprise_linux:7.2:GA:hypervisor"
HOME_URL="https://www.redhat.com/"
BUG_REPORT_URL="https://bugzilla.redhat.com/"

# FIXME
REDHAT_BUGZILLA_PRODUCT="Red Hat Virtualization"
REDHAT_BUGZILLA_PRODUCT_VERSION=7.2
REDHAT_SUPPORT_PRODUCT="Red Hat Virtualization"
REDHAT_SUPPORT_PRODUCT_VERSION=7.2

# cat /etc/system-release
Red Hat Enterprise Linux release 7.2

# cat /etc/system-release-cpe
cpe:/o:redhat:enterprise_linux:7.2:ga:hypervisor

# cat /etc/redhat-release
Red Hat Enterprise Linux release 7.2


GUI: Hosts - General - Software

OS Version:
RHEL - 4.0 - 1.el7
OS Description:
Red Hat Virtualization Host 4.0 (el7.2)

Ok, it's wrong  in the UI, but coirrect in the files.
vdsm must be pulling it from the incorrect source.

Eli, can you clarify where vdsm is pulling the version from? I currently can't see swhere it is getting the "4.0" version from.

(In reply to Fabian Deutsch from comment #15)
> Ok, it's wrong  in the UI, but coirrect in the files.
> vdsm must be pulling it from the incorrect source.
> 
> Eli, can you clarify where vdsm is pulling the version from? I currently
> can't see swhere it is getting the "4.0" version from.

Fabian, VDSM uses the same code as for RHEL, so we are getting version and release from querying basename package.

How do you identify the basename package?

The assumption was that one of /etc/os-release, /etc/system-release, /etc/system-release-cpe and /etc/redhat-release was directly or indirectly used.

Please take a look at 

https://gerrit.ovirt.org/gitweb?p=vdsm.git;a=blob;f=lib/vdsm/osinfo.py;h=3c6ab3bd99e0c7a127f551c38b4697bf79527fff;hb=refs/heads/master#l164

Okay, this explains it.

So not the contents of this file is used, but rather the version/release of the owner of that file.

Which is a different release file on Node.

In the UI it is not to much of an issue.
But I suppose tha this version/released is used for some decision, i.e. allowed update?

I've opened bug 1368364 to track the incorrect reporting.

Ths bug is actually verified, because pretty name is reported correctly.

Discussed over BJ, clearing needinfo