2

I am trying to utilize AutoPilotBeta to deploy OpenStack on 8 Dell servers. The servers are a mix of R610, R415, R815, and R310 models. Some of these servers report DVD, CD, and NVRam as Disk in response to the lshw command, but do not report size. I have gotten to the point in the process where I have a MAAS configured and all of the servers are commissioned. When I execute the "openstack-install", it completes successfully. When I go to the https://<my-server-ip>/account/standalone/openstack/ page, I am able to login and I have all green checkmarks. I click and I get all of the options to configure my cloud except the physical zones. when I check the error log on the server that is running Landscape, I see the following:

File "/opt/canonical/landscape/canonical/landscape/model/openstack/jobs.py", line 260, in run
disk_size = sum(int(disk["size"]) for disk in disks)
File "/opt/canonical/landscape/canonical/landscape/model/openstack/jobs.py", line 260, in <genexpr>
disk_size = sum(int(disk["size"]) for disk in disks)
exceptions.TypeError: int() argument must be a string or a number, not 'NoneType'

I tried querying the the API for the same node the threw the error ... I am pretty sure this is the section that caused it:

  <node id="disk:0" claimed="true" class="disk" handle="SCSI:00:00:00:00">
   <description>SCSI Disk</description>
   <physid>0</physid>
   <businfo>scsi@0:0.0.0</businfo>
   <logicalname>/dev/sdc</logicalname>
   <dev>8:32</dev>
   <configuration>
    <setting id="sectorsize" value="512" />
   </configuration>
  </node>

No mention of size anywhere ... since it is a driver for USB storage that is not plugged in, so it has no size ...

Is there a known work around for this issue?

Improve this question

1 Answer 1

Reset to default
2

UPDATE: 06 Apr 2015: As of this date, the official instructions are now installing LDS 15.01 that gets around this problem.

I believe you are running into github issue 374 which has been addressed in LDS 15.01.

Notice in the release notes that upgrading the juju deployed LDS is not directly supported (though it can be done if you follow the non-quickstart upgrade steps closely and understand a bit about juju).

Instead, we recommend installing fresh. You can follow these steps for something very quick (after releasing the resources already allocated in MAAS):

sudo apt-get install juju-quickstart
juju quickstart u/landscape/landscape-dense-maas/12

# After it's done... get the ip address of the apache2 node:
juju status apache2

# And connect to that in your web browser
Improve this answer
4
  • This was exactly the information that I have been looking for.
    – Pat E
    Apr 5, 2015 at 9:54
  • Just for the sake of anybody else who may run into this issue and, like me, is unable to find the way around it. I made it to step 5 of the official instructions and then was stopped and there was no clear path to a solution. I needed to copy the credentials from the ~/.juju/environments.yaml file, then perform a "sudo openstack-install -u" to clean out the failed install, then I had to bootstrap juju, following the instructions provided for doing that.Then I followed the instructions above.
    – Pat E
    Apr 5, 2015 at 10:07
  • It tried to throw me into a text-based browser session with the juju admin interface, I just connected to it an accepted what it was telling me. After that, I issues a sudo juju status apache2 (as described above). That gave me the IP address of the landscape server. I connected to it and was able to add my nodes to my physical zone! Once again, Thank you!
    – Pat E
    Apr 5, 2015 at 10:08
  • I just pushed new packages to stable ppa that will have the latest Autopilot from here on out
    – battlemidget
    Apr 6, 2015 at 13:00

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .