maas-devel team mailing list archive

[Andreas Hasenack <andreas@xxxxxxxxxxxxx>]

On Thu, Jul 3, 2014 at 9:44 AM, Andreas Hasenack <andreas@xxxxxxxxxxxxx>
wrote:

> On Thu, Jul 3, 2014 at 9:06 AM, Gavin Panella <gavin.panella@xxxxxxxxxxxxx
> > wrote:
>
>> On 2 July 2014 20:08, Andreas Hasenack <andreas@xxxxxxxxxxxxx> wrote:
>> > I upgraded from 1.5.2 to this release above and noticed that my boot
>> images
>> > were gone. Is that expected?
>>
>> I wouldn't expect that. The maas-cluster-controller.postrm script will
>> remove the whole of /var/lib/maas when purging, so is there any chance
>> that 1.5.2 got purged on the way to 1.6?
>>
>>
>>
>
> The only odd thing in this upgrade, and happens everytime, is that the
> rabbitmq and maas host ips get replaced by an IP (192.168.122.6) that I
> don't have. I don't know where it comes from, so in the end the services
> are not up and I have to fix it manually in these two files:
>
> root@maas:/etc/maas# grep 192 *
> maas_local_celeryconfig.py:BROKER_URL = 'amqp://
> maas_workers:Pd8Lj1mvYqPW618XoIhZ@192.168.122.6:5672//maas_workers'
> maas_local_settings.py:DEFAULT_MAAS_URL = "http://192.168.122.6/MAAS";
>
>
> The whole upgrade is here: http://pastebin.ubuntu.com/7741746/
>
> After fixing that and noticing my images were gone, I clicked the import
> button, and now I have 129 (!!) images, including arm ones :) My disk even
> got full before that finished, which was already an indication that it was
> importing more than I expected :)
>
>
Another issue I had was a dpkg error while importing images:


[2014-07-02 20:50:32,017: ERROR/Worker-2] import_boot_images: Command `dpkg
-x /tmp/maas-XT_xbX/shim-signed_1.6+0.4-0ubuntu4_amd64.deb
/tmp/maas-XT_xbX` returned non-zero exit status 2:

My /tmp was lke this:
 overflow        1.0M   28K  996K   3% /tmp

Don't remember what set that up. 1MB is definitely small.

For the moment I just umounted it and let /tmp be on the rootfs, then the
rest worked.