Tag Info

Hot answers tagged

5

I believe you will have to redeploy the service on that machine. Even in MAAS, it's usually best to recommission the node if its hardware changed.


3

Someone (could have been me) had done something clever trying to proxy redirect web services using the same apache which is servicing maas. This was in /etc/apache2/sites-enabled This breaks /MAAS <VirtualHost *:80 > ServerName example.com ServerAlias www.example.com ProxyPass / http://10.0.0.151:6543/ </VirtualHost > Be sure to include ...


2

So you can model fault tolerance in both juju and MAAS. With MAAS you can set up multiple regions, while regions can be anything you want, they typically refer to an entire isolated power+networked area. (Imagine a region being a datacenter). If you configure MAAS in this way using it with juju becomes easy. After running juju bootstrap you can run juju ...


2

You could use the local provider to experiment with (https://jujucharms.com/docs/1.20/config-local), but not for production services. It will deploy services in LXC containers on your local machine. Other than that, there is a manual provider (https://jujucharms.com/docs/1.20/config-manual), where you tell juju to use specific machines.


2

Try to login on the machines that are stuck in "allocating" and see what's going on. Try juju ssh <unit-name>, or just juju ssh N where "N" is the machine id. Also, the debug-log command you used will by default only show the last few lines, like tail(1). Try juju debug-log --replay, that will give you the full output. But I think your best bet is to ...


2

MAAS will make sure /you/ have access to each node by adding the key that Juju tells it (and that key is only your public-key). Units do not have SSH access to themselves by design (think of the security implications!). If you wanted to make it so that all units or services can access each other you need to have each machines generate an SSH key for the ...


2

1) Could be a possible bug, the MAAS UI should support multiple entries. 2) You'll need to make your changes within /etc/bind/maas/named.conf* as those files are included into the top level named options.


2

Ok, a few things: a) The syntax for no_proxy is an IP or a hostname, separated by commas, not a URL. So yours should have been just the IP. That being said, I don't think you need (or should) specify the proxy itself in there. Take a look at https://jujucharms.com/docs/1.22/howto-proxies b) I think you are missing a port in the http-proxy variable. Or did ...


2

The warning is unfortunate. Looks like anything you set for the first time will yield that loud WARNING. Checkout this example: andreas@nsn7:~$ juju unset-env http-proxy andreas@nsn7:~$ andreas@nsn7:~$ juju unset-env http-proxy WARNING key "http-proxy" is not defined in the current environment configuration: possible misspelling andreas@nsn7:~$ ...


2

If you use the latest Landscape Autopilot, LDS 15.01, then you need at least: one machine for MAAS one machine for LDS 3 machines for OpenStack Details about the networking setup can be found here: https://insights.ubuntu.com/2015/04/10/maas-network-layouts-for-the-landscape-autopilot/ Or at this askubuntu q&a: How should I setup MAAS so that it ...


1

While it is "stuck" there, can you ssh into 172.16.1.141 (the address from your screenshot, it will vary between attempts) as ubuntu, from the same machine where you are bootstrapping from? I've seen this before and there was a stuck dpkg or apt-get process on that node as a result of an incorrect MTU. Meaning, ssh worked, but as soon as a transfer of ...


1

Note: Clicking resolve in the GUI as you did you're actually telling Juju that the issue has been resolved not telling it to resolve the issue. It's hard to debug why Wordpress isn't working in this case without actually poking at it but the failure to add the new unit may have put the charm in a broken state which hasn't actually been resolved (see ...


1

we haven't yet finalized the 1.8 release plan but plan to have a beta out this month. Are there any specific features in 1.8 that you care about that are not in 1.7? We could look into backporting if necessary.


1

Did you change the interfaces on these two nodes after commissioning? If you did, can you try commissioning them again in their current state and then try to bootstrap with juju manually? Instructions on how to configure juju with MAAS can be found here: https://jujucharms.com/docs/1.20/config-maas


1

You use deploy when the service doesn't exist yet. You use add-unit when the service already exists.


1

In 14.10 MAAS now handles the importing of boot images. You know longer need to run that script. On the new "Images" page you can start the import process. You can also do this from the command line. Once you have logged in using the command line. [Documentation for logging over the command line is here].1 maas boot-resources import You can then monitor ...


1

I suspect your MAAS node cannot reach out to the internet to install the necessary software. The charm-helpers PPA is being used by the node, not your local machine.


1

Proxy settings for juju are explained in this question/answer: juju http-proxy and no_proxy settings


1

In answer to your general question, the cloud-init package is responsible for generating host keys during first boot on cloud images. In your more specific case, I don't think it currently generates ed25519 keys, but this hasn't been known to cause any issues for any users. If it does cause a problem for you, you should note your use case in more detail in ...


1

This is already happening in parallel. I was not checking the times correctly.


1

If, in maas, 10.10.10.102 is set as your gateway, this would be expected behavior. At least it would be by me, having been bitten by this king of thing. I renamed the machine I was use for juju bootstrap to jujuBS, so try this, which bring everything you need local to the maas server: mkdir ~./.juju/sync-tools juju sync-tools –e maas ...


1

I had a problem similar to this, but not exactly the same. I tracked it down to having set "Static IP low" and "Static IP high" for the network the cluster controller was using. Once I cleared the fields, the problem went away and it booted correctly. Or, more correctly, I went on to other problems.


1

I also had this problem. The root of the problem for me was that I had no internet access and needed to to do a couple of things. IIRC, this particular this was being caused by the time of the server and target machine being off by too large an amount of time. It was not actually stuck, but the time out was huge, maybe 30 minutes. I set my server as an ...



Only top voted, non community-wiki answers of a minimum length are eligible