Reprovision of nodes of new Fedora-cloud

[Miroslav Suchý]

Stephan,
can you please reprovision fed-cloud09-15? I.e. both controller and all
compute nodes.

I'm ccing mailing list so if you experimenting with FC, you know why
your VM are gone. There should be none production VM.

It may be final version, but I'm not sure, because I made some changes
to packstack, which I could not test right now. So it may be final or I
may request one more reprovision.

Once declared final, I plan to do Fedora classroom for everybody who
want to learn how it is set up, what you can do and what you should not do.

Mirek

[Stephen John Smoogen]

On 20 April 2015 at 07:07, Miroslav Suchy  wrote:
> Stephan,
> can you please reprovision fed-cloud09-15? I.e. both controller and all
> compute nodes.

I will start on this project in a short while. My apologies for the
delay... this got stuck in a folder for autoemails for some reason.
> I'm ccing mailing list so if you experimenting with FC, you know why
> your VM are gone. There should be none production VM.
>
> It may be final version, but I'm not sure, because I made some changes
> to packstack, which I could not test right now. So it may be final or I
> may request one more reprovision.
>
> Once declared final, I plan to do Fedora classroom for everybody who
> want to learn how it is set up, what you can do and what you should not do.
>
> Mirek
> _______________________________________________
> infrastructure mailing list
> [e-mail]
> https://admin.fedoraproject.org/mailman/listinfo/infrastructure

[Stephen John Smoogen]

On 20 April 2015 at 18:18, Stephen John Smoogen  wrote:
> On 20 April 2015 at 07:07, Miroslav Suchy  wrote:
> > Stephan,
> > can you please reprovision fed-cloud09-15? I.e. both controller and all
> > compute nodes.
>
> I will start on this project in a short while. My apologies for the
> delay... this got stuck in a folder for autoemails for some reason.

I have run into a crashing bug on rebuilding boxes..
fed-cloud09, fed-cloud10, fed-cloud11 are rebuilt
fed-cloud12 -> fed-cloud15 are not.

I will finish fed-cloud12 -> fed-cloud15 later 'today' when I get some
feedback on crasher.

[Stephen John Smoogen]

On 20 April 2015 at 07:07, Miroslav Suchy  wrote:
> Stephan,
> can you please reprovision fed-cloud09-15? I.e. both controller and all
> compute nodes.
>
> I'm ccing mailing list so if you experimenting with FC, you know why
> your VM are gone. There should be none production VM.
>
> It may be final version, but I'm not sure, because I made some changes
> to packstack, which I could not test right now. So it may be final or I
> may request one more reprovision.
>
> Once declared final, I plan to do Fedora classroom for everybody who
> want to learn how it is set up, what you can do and what you should not do.
>
> Mirek
> _______________________________________________
> infrastructure mailing list
> [e-mail]
> https://admin.fedoraproject.org/mailman/listinfo/infrastructure

Systems have been rebuilt. ansible on fed-cloud09 currently dies at

TASK: [command vgrename vg_guests cinder-volumes]
*****************************
failed: [fed-cloud09.cloud.fedoraproject.org] => {"changed": true, "cmd":
["vgrename", "vg_guests", "cinder-volumes"], "delta": "0:00:00.040102",
"end": "2015-04-21 19:25:44.696825", "rc": 5, "start": "2015-04-21
19:25:44.656723", "warnings": []}
stderr:   New volume group "cinder-volumes" already exists
...ignoring

TASK: [lvg vg=cinder-volumes pvs=/dev/md127 pesize=32 vg_options=""]
**********
ok: [fed-cloud09.cloud.fedoraproject.org]

TASK: [Create logical volume for Swift]
***************************************
ok: [fed-cloud09.cloud.fedoraproject.org]

TASK: [Create FS on Swift storage]
********************************************
ok: [fed-cloud09.cloud.fedoraproject.org]

TASK: [template src={{ files }}/fedora-cloud/hosts dest=/etc/hosts
owner=root mode=0644] ***
ok: [fed-cloud09.cloud.fedoraproject.org]

TASK: [stat path=/etc/packstack_sucessfully_finished]
*************************
ok: [fed-cloud09.cloud.fedoraproject.org]

TASK: [service name=NetworkManager state=stopped enabled=no]
******************
ok: [fed-cloud09.cloud.fedoraproject.org]

TASK: [service name=network state=started enabled=yes]
************************
failed: [fed-cloud09.cloud.fedoraproject.org] => {"failed": true}
msg: Job for network.service failed. See 'systemctl status network.service'
and 'journalctl -xn' for details.

FATAL: all hosts have already failed -- aborting

=====

I ran it twice in case it was a "we needed to run it once and then again to
get past this" type error.

[Stephen John Smoogen]

Further update:

fed-cloud10 through fed-cloud15 have had the openstack playbook run on
them. For some reason the boxes fed-cloud12 -> fed-cloud15 would not run
correctly because something either deleted the eth1 file before the
playbook was run or during the run. Not sure what is causing it to work on
fed-cloud09->fed-cloud10