Restoring admin to a healthy state

SUSE OpenStack Cloud
1

Hi

This is my first post here, i have some troubles with crowbar, i try to add a new external network to an existing SOC7 instalation with the following steeps:

  1. edit crowbar network proposal with “crowbar network proposal edit default”

  2. Add the new network on the networks section of the json following these instructions: https://www.suse.com/documentation/suse-openstack-cloud-7/book_cloud_deploy/data/sec_setup_multi_ext_networks.html

  3. deploy de modified conduict with “crowbar network proposal commit default”.

The admin server starts to try to deploy de proposal, but stops when try to reach de admin server.

Now the network barclamp is broken and i dont have any idea how to get the admin server to a healthy state again in order to try again to redeploy de new network.

Thanks in advice for the help.

2

On 05/10/17 00:04, ehernandezc wrote:
[color=blue]

This is my first post here, i have some troubles with crowbar, i try to
add a new external network to an existing SOC7 instalation with the
following steeps:

  1. edit crowbar network proposal with “crowbar network proposal edit
    default”

  2. Add the new network on the networks section of the json following
    these instructions:
    https://www.suse.com/documentation/suse-openstack-cloud-7/book_cloud_deploy/data/sec_setup_multi_ext_networks.html

  3. deploy de modified conduict with “crowbar network proposal commit
    default”.

The admin server starts to try to deploy de proposal, but stops when try
to reach de admin server.[/color]

What change(s) are you making to the network proposal? After making a
simple change (mapping a different physical interface to a logical
interface) I am able to commit the edited proposal.

SUSE’s documentation advises to edit /etc/crowbar/network.json file by
hand before installing SUSE OpenStack Cloud and notes that once
installed you cannot change the network setup. What’s not clear is
whether this also applies to setting up multiple external networks and
modifying the Neutron Barclamp - I’ll report this.

When making changes using “crowbar network proposal edit default” with a
customised (by hand) network.json file in /etc/network I don’t then see
those changes reflected in the network.json file. Whilst it seems to
read in my custom config this makes me think using “crowbar network
proposal edit default” is not supported.
[color=blue]

Now the network barclamp is broken and i dont have any idea how to get
the admin server to a healthy state again in order to try again to
redeploy de new network.

Thanks in advice for the help.[/color]

Did you backup the Admin node before making network changes?

Do you have nodes set up with services deployed?

HTH.

Simon
SUSE Knowledge Partner

If you find this post helpful and are logged into the web interface,
please show your appreciation and click on the star below. Thanks.

3

Hi,

I wanted to reply to this a month ago, but my account was deactivated accidentally :smiley:

First of all, the network.json is for initial configuration of the servers, this configuration should indeed be correct and final, otherwise a reinstallation will be necessary.
Adding more external networks after the cloud has been already deployed works fine. You actually don’t need to edit the network.json for that, but you need to edit the network barclamp directly (in raw mode) and the redeploy it. This will add the new external network to the database. After that, you need to edit the neutron barclamp in raw mode and add the attribute “additional_external_networks” with respective values for your new network(s), redeploy neutron barclamp. Then you need to create that network with neutron net-create command, but please note there’s a network name missing in the command example in the docs. The command should look like this (for linuxbridge):

sudo neutron net-create --router:external True --provider:physical_network physnet1 \\ --provider:network_type vlan --provider:segmentation_id 600 public2

That’s this part.
Resetting your admin server to a healthy state can be tricky. It saves you a lot of trouble if you use a virtual machine for admin server and take snapshots, at least one right after the cloud has been installed on admin server and before any changes are made.
I also experienced failing connection to the admin server in some cases, it could be that your network changes are not valid, did you check them with the json-validator?
Try restarting crowbar.service before redeploying the barclamp, are all services like chef up and running?

And then there’s the hard way:
If you don’t have any control or compute nodes registered yet, try a clean install of the cloud. But I have had some troubles with simply reinstalling the addon etc., I had to wipe everything clean by removing all packages related to crowbar, chef, couchdb, ruby etc. But make sure to keep a copy of your network.json before deleting anything! Then I deleted all remaining files and directories manually after deinstallation, then reinstall the cloud.

Since you didn’t reply yet I assume you either fixed it already or gave up? :stuck_out_tongue: I’ll leave this here anyway, maybe this helps someone else, too.