[Users] Trouble upgrading

Fri Nov 22 12:53:21 UTC 2013

On 11/22/2013 06:54 AM, Kristaps wrote:
> Bob Doolittle <bob at ...> writes:
>
>>
>> On 11/21/2013 12:57 PM, Itamar Heim wrote:
>>> On 11/21/2013 07:38 PM, Bob Doolittle wrote:
>>>> On 11/21/2013 12:00 PM, Itamar Heim wrote:
>>>>> On 11/21/2013 06:32 PM, Bob Doolittle wrote:
>>>>>> Yay!
>>>>>>
>>>>>> Congratulations to all of the oVirt team.
>>>>>>
>>>>>> I am having trouble locating upgrade instructions, however. There's
>>>>>> nothing in the release notes.
>>>>>>
>>>>>> I discovered through trial-and-error that running "engine-setup"
> again
>>>>>> handles upgrade of the Engine.
>>>>>>
>>>>>> But I don't know how to upgrade my RH 6.4 KVM host. When I try to run
>>>>>> "yum update" it fails due to dependency errors notably in:
>>>>>> glusterfs
>>>>>> qemu
>>>>>> vdsm
>>>>>>
>>>>> can you please include the yum output log?
>>>> Sorry yes. Attached are three outputs. First my repolist. Then yum
>>>> update with no options, and yum update with --skip-broken (so you can
>>>> see the lurking issues masked by the first failure).
>>>>
>>>> It appears that the glusterfs issue is due to a newer version offered
> by
>>>> rhel-x86_64-server-6: glusterfs-3.4.0.36rhs-1.el6.x86_64. This has
>>>> dependencies which conflict with the version provided by ovirt-stable
>>>> (and glusterfs-epel).
>>>>
>>>> It looks like there was a large update to the Red Hat repo last night
>>>> that has caused conflicts with ovirt-stable.
>>>>
>>>> Note I can get it to the point where it's willing to install packages
> if
>>>> I specify --skip-broken and also --exclude vdsm-python-4.12.1, but
>>>> that's a bit too scary for me without checking first.
>>> oh, you may caught rhel repo in middle of a refresh - try again in a
>>> few hours first, but could be the refresh may cause an issue we'd need
>>> to resolve.
>> Not sure why I'm the only person reporting this (am I the only person
>> who runs RH 6 on their KVM hosts who has tried upgrading??), but the
>> problem has not resolved.
>>
>> What's the next step? Does it need a bug opened?
>>
>> Let me know if I can provide any more information.
>>
>> Thanks,
>>       Bob
>>
>>>> Thanks,
>>>>       Bob
>>>>
>>>>>> and also to a multilib version error due to vdsm-python 4.12 and
> 4.13.
>>>>>> What's the proper upgrade procedure for a Host?
>>>>>>
>>>>>> Thanks,
>>>>>>       Bob
>>>>>>
>>>>>> On 11/21/2013 10:43 AM, Kiril Nesenko wrote:
>>>>>>> The oVirt development team is very happy to announce the general
>>>>>>> availability of oVirt 3.3.1 as of November 21th 2013. This release
>>>>>>> solidifies oVirt as a leading KVM management application, and open
>>>>>>> source alternative to VMware vSphere.
>>>>>>>
>>>>>>> oVirt is available now for Fedora 19 and Red Hat Enterprise Linux
> 6.4
>>>>>>> (or similar).
>>>>>>>
>>>>>>> See release notes [1] for a list of the new features and bug fixed.
>>>>>>>
>>>>>>> [1] http://www.ovirt.org/OVirt_3.3.1_release_notes
>>>>>>>
>>>>>>> - Kiril
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Users mailing list
>>>>>>> Users at ...
>>>>>>> http://lists.ovirt.org/mailman/listinfo/users
>>>>>> _______________________________________________
>>>>>> Users mailing list
>>>>>> Users at ...
>>>>>> http://lists.ovirt.org/mailman/listinfo/users
> I'am experiencing same exact issue with upgrade on RHEL 6.4 and vdsm for
> oVirt installed.
>
> However, I did a little searching, and for me it seems it is like this:
> vds requires these glsuterfs packages to have the latest version:
>
> glusterfs
> glusterfs-api
> glusterfs-cli
> glusterfs-fuse
> glusterfs-libs
>
> In RHEL repository are these packages are with version 3.4.0.36rhs-
> 1.el6.x86_64 EXCEPT for glusterfs-cli, which RHEL repo doesn't have at all.
>
> oVirt repo has all those packages with version 3.4.0-8.el6.x86_64, which is
> older than RHEL repo packages.
>
> Hence, there is conflict. During vdsm update or even installation on RHEL
> 6.4 it tries to install the latest of all those packages. Looks likes this:
>
> glusterfs-3.4.0.36rhs-1.el6.x86_64
> glusterfs-api-3.4.0.36rhs-1.el6.x86_64
> glusterfs-fuse-3.4.0.36rhs-1.el6.x86_64
>     requires: glusterfs-libs-3.4.0.36rhs-1.el6.x86_64
>
> glusterfs-cli-3.4.0-8.el6.x86_64
>     requires: glusterfs-libs-3.4.0-8.el6.x86_64
>
> While this is not fixed, in my opinion, ovirt installation from repository
> on RHEL 6.4 is essentially broken.
>
> _______________________________________________
> Users mailing list
> Users at ovirt.org
> http://lists.ovirt.org/mailman/listinfo/users

Thank you for confirming that this is indeed an issue. I was beginning 
to think it was just me somehow. I have opened 
https://bugzilla.redhat.com/show_bug.cgi?id=1033587 to track this.

I agree it seems to be a very big issue for RHEL 6 users. And we can't 
simply uninstall glusterfs-cli, as Vijay suggested, because it is 
required by vdsm. So I'm not seeing any easy workaround at the moment.

I'm wondering if the qemu issue I was when I tried --skip-broken is just 
an artifact of this, or some other issue entirely. We won't know until 
this glusterfs-cli issue is resolved.

-Bob