Tag: OpenStack

"Fishing line and bobbin stuck on tree at Douthat State Park" by "Virginia State Parks" on Flickr

Note to self: Linux shell scripts don’t cope well with combined CRLF + LF files… Especially in User-Data / Custom Data / Cloud-Init scripts

JonTheNiceGuy 4 Comments

This one is more a nudge to myself. On several occasions when building Infrastructure As Code (IAC), I split out a code sections into one or more files, for readability and reusability purposes. What I tended to do, and this was more apparent with the Linux builds than the Windows builds, was to forget to set the line terminator from CRLF to LF.

While this doesn’t really impact Windows builds too much (they’re kinda designed to support people being idiots with line endings now), Linux still really struggles with CRLF endings, and you’ll only see when you’ve broken this because you’ll completely fail to run any of the user-data script.

How do you determine this is your problem? Well, actually it’s a bit tricky, as neither cat, less, more or nano spot this issue. The only two things I found that identified it were file and vi.

The first part of the combined file with mixed line endings. This part has LF termination.
The second part of the combined file with mixed line endings. This part has CRLF termination.
What happens when we cat these two parts into one file? A file with CRLF, LF line terminators obviously!
What the combined file looks like in Vi. Note the blue ^M at the ends of the lines.

So, how to fix this? Assuming you’re using Visual Studio Code;

A failed line-ending clue in Visual Studio Code

You’ll notice this line showing “CRLF” in the status bar at the bottom of Code. Click on that, which brings up a discrete box near the top, as follows:

Oh no, it’s set to “CRLF”. That’s not what we want!

Selecting LF in that box changes the line feeds into LF for this file, but it’s not saved. Make sure you save this file before you re-run your terraform script!

Notice, we’re now using LF endings, but the file isn’t saved.

Fantastic! It’s all worked!

In Nano, I’ve opened the part with the invalid line endings.

Oh no! We have a “DOS Format” file. Quick, let’s fix it!

To fix this, we need to write the file out. Hit Ctrl+O. This tells us that we’re in DOS Format, and also gives us the keyboard combination to toggle “DOS Format” off – it’s Alt+D (In Unix/Linux world, the Alt key is referred to as the Meta key – hence M not A).

This is how we fix things

So, after hitting Alt+D, the “File Name to write” line changes, see below:

Yey, no pesky “DOS Format” warning here!

Using either editor (or any others, if you know how to solve line ending issues in other editors), you still need to combine your script back together before you can run it, so… do that, and your file will be fine to run! Good luck!

Featured image is “Fishing line and bobbin stuck on tree at Douthat State Park” by “Virginia State Parks” on Flickr and is released under a CC-BY license.

Creating OpenStack “Allowed Address Pairs” for Clusters with Ansible

JonTheNiceGuy Leave a comment

This post came about after a couple of hours of iterations, so I can’t necessarily quote all the sources I worked from, but I’ll do my best!

In OpenStack (particularly in the “Kilo” release I’m working with), if you have a networking device that will pass traffic on behalf of something else (e.g. Firewall, IDS, Router, Transparent Proxy) you need to tell the virtual NIC that the interface is allowed to pass traffic for other IP addresses, as OpenStack applies by default a “Same Origin” firewall rule to the interface. Defining this in OpenStack is more complex than it could be, because for some reason, you can’t define 0.0.0.0/0 as this allowed address pair, so instead you have to define 0.0.0.0/1 and 128.0.0.0/1.

Here’s how you define those allowed address pairs (note, this assumes you’ve got some scaffolding in place to define things like “network_appliance”):

allowed_address_pairs: "{% if (item.0.network_appliance|default('false')|lower() == 'true') or (item.1.network_appliance|default('false')|lower() == 'true') %}[{'ip_address': '0.0.0.0/1'}, {'ip_address': '128.0.0.0/1'}]{% else %}{{ item.0.allowed_address_pairs|default(omit) }}{% endif %}"

OK, so we’ve defined the allowed address pairs! We can pass traffic across our firewall. But (and there’s always a but), the product I’m working with at the moment has a floating MAC address in a cluster, when you define an HA pair. They have a standard schedule for how each port’s floating MAC is assigned… so here’s what I’ve ended up with (and yes, I know it’s a mess!)

allowed_address_pairs: "{% if (item.0.network_appliance|default('false')|lower() == 'true') or (item.1.network_appliance|default('false')|lower() == 'true') %}[{'ip_address': '0.0.0.0/1'},{'ip_address': '128.0.0.0/1'}{% if item.0.ha is defined and item.0.ha != '' %}{% for vdom in range(0,40, 10) %},{'ip_address': '0.0.0.0/1','mac_address': '{{ item.0.floating_mac_prefix|default(item.0.image.floating_mac_prefix|default(floating_mac_prefix)) }}:{% if item.0.ha.group_id|default(0) < 16 %}0{% endif %}{{ '%0x' | format(item.0.ha.group_id|default(0)|int) }}:{% if vdom+(item.1.interface|default('1')|replace('port', '')|int)-1 < 16 %}0{% endif %}{{ '%0x' | format(vdom+(item.1.interface|default('1')|replace('port', '')|int)-1) }}'}, {'ip_address': '128.0.0.0/1','mac_address': '{{ item.0.floating_mac_prefix|default(item.0.image.floating_mac_prefix|default(floating_mac_prefix)) }}:{% if item.0.ha.group_id|default(0) < 16 %}0{% endif %}{{ '%0x' | format(item.0.ha.group_id|default(0)|int) }}:{% if vdom+(item.1.interface|default('0')|replace('port', '')|int)-1 < 16 %}0{% endif %}{{ '%0x' | format(vdom+(item.1.interface|default('1')|replace('port', '')|int)-1) }}'}{% endfor %}{% endif %}]{% else %}{{ item.0.allowed_address_pairs|default(omit) }}{% endif %}"

Let's break this down a bit. The vendor says that each port gets a standard prefix, (e.g. DE:CA:FB:AD) then the penultimate octet is the "Cluster ID" number in hex, and then the last octet is the sum of the port number (zero-indexed) added to a VDOM number, which increments in 10's. We're only allowed to assign 10 "allowed address pairs" to an interface, so I've got the two originals (which are assigned to "whatever" the defined mac address is of the interface), and four passes around. Other vendors (e.g. this one) do things differently, so I'll probably need to revisit this once I know how the next one (and the next one... etc.) works!

So, we have here a few parts to make that happen.

The penultimate octet, which is the group ID in hex needs to be two hex digits long, and without adding more python modules to our default machines, we can't use a "pad" filter (to add 0's to the beginning of the mac octets), so we do that by hand:

{% if item.0.ha.group_id|default(0) < 16 %}0{% endif %}

And here's how to convert the group ID into a hex number:

{{ '%0x' | format(item.0.ha.group_id|default(0)|int) }}

Then the next octet is the sum of the VDOM and PortID. First we need to loop around the VDOMs. We don't always know whether we're going to be adding VDOMs until after deployment has started, so here we will assume we've got 3 VDOMs (plus VDOM "0" for management) as it doesn't really matter if we don't end up using them. We create the vdom variable like this:

{% for vdom in range(0, 40, 10) %} STUFF {% endfor %}

We need to put the actual port ID in there too. As we're using a with_subelement loop we can't create an increment, but what we can do is ensure we're recording the interface number. This only works here because the vendor has a sequential port number (port1, port2, etc). We'll need to experiment further with other vendors! So, here's how we're doing this. We already know how to create a hex number, but we do need to use some other Jinja2 filters here:

{{ '%0x' | format(vdom+(item.1.interface|default('1')|replace('port', '')|int)-1) }}

Let's pull this apart a bit further. item.1.interface is the name of the interface, and if it doesn't exist (using the |default('1') part) we replace it with the string "1". So, let's replace that variable with a "normal" value.

{{ '%0x' | format(vdom+("port1"|replace('port', '')|int)-1) }}

Next, we need to remove the word "port" from the string "port1" to make it just "1", so we use the replace filter to strip part of that value out. Let's do that:

{{ '%0x' | format(vdom+("1"|int)-1) }}

After that, we need to turn the string "1" into the literal number 1:

{{ '%0x' | format(vdom+1-1) }}

We loop through vdom several times, but let's pick one instance of that at random - 30 (the fourth iteration of the vdom for-loop):

{{ '%0x' | format(30+1-1) }}

And then we resolve the maths:

{{ '%0x' | format(30) }}

And then the |format(30) turns the '%0x' into the value "1e"

Assuming the vendor prefix is, as I mentioned, 'de:ca:fb:ad:' and the cluster ID is 0, this gives us the following resulting allowed address pairs:

[
{"ip_address": "0.0.0.0/1"},
{"ip_address": "128.0.0.0/1"},
{"ip_address": "0.0.0.0/1", "mac_address": "de:ca:fb:ad:00:00"},
{"ip_address": "128.0.0.0/1", "mac_address": "de:ca:fb:ad:00:00"},
{"ip_address": "0.0.0.0/1", "mac_address": "de:ca:fb:ad:00:0a"},
{"ip_address": "128.0.0.0/1", "mac_address": "de:ca:fb:ad:00:0a"},
{"ip_address": "0.0.0.0/1", "mac_address": "de:ca:fb:ad:00:14"},
{"ip_address": "128.0.0.0/1", "mac_address": "de:ca:fb:ad:00:14"},
{"ip_address": "0.0.0.0/1", "mac_address": "de:ca:fb:ad:00:1e"},
{"ip_address": "128.0.0.0/1", "mac_address": "de:ca:fb:ad:00:1e"}
]

I hope this has helped you!

Sources of information:

Getting a PEM file from your OpenSSH Private Key

JonTheNiceGuy Leave a comment

At work, the system used to get a Windows Administrator password in our OpenStack based system (K5) is derived from the SSH Public Key recorded in the system.

It’s really easy to use, and can be found here: https://decrypt-win-passwd.uk-1.cf-app.net

There is one downside to this though – the application needs the private key to be supplied to it (it’s OK, you regularly rotate your SSH private keys… right??) in PEM format… Now, if you’re any sort of sensible SSH user, you’ve used either OpenSSH’s ssh-keygen command, or PuTTY’s puttygen command… neither of which produce a PEM format key.

So, you need to convert it. After a bit of proding and poking, I found this command

openssl rsa -outform PEM -in ~/.ssh/id_rsa -out ~/.ssh/id_rsa.pem

Like the last post, this is more for me to find stuff in the future, but… if he helps someone else, so much the better!!

Today I learned… that you can look at the “cloud-init” files on your target server…

JonTheNiceGuy Leave a comment

Today I have been debugging why my Cloud-init scripts weren’t triggering on my Openstack environment.

I realised that something was wrong when I tried to use the noVNC console[1] with a password I’d set… no luck. So, next I ran a command to review the console logs[2], and saw a message (now, sadly, long gone – so I can’t even include it here!) suggesting there was an issue parsing my YAML file. Uh oh!

I’m using Ansible’s os_server module, and using templates to complete the userdata field, which in turn gets populated as cloud-init scripts…. and so clearly I had two ways to debug this – prefix my ansible playbook with a few debug commands, but then that can get messy… OR SSH into the box, and look through the logs. I knew I could SSH in, so the cloud-init had partially fired, but it just wasn’t parsing what I’d submitted. I had a quick look around, and found a post which mentioned debugging cloud-init. This mentioned that there’s a path (/var/lib/cloud/instances/$UUID/) you can mess around in, to remove some files to “fool” cloud-init into thinking it’s not been run… but, I reasoned, why not just see what’s there.

And in there, was the motherlode – user-data.txt…. bingo.

In the jinja2 template I was using to populate the userdata, I’d referenced another file, again using a template. It looks like that template needs an extra line at the end, otherwise, it all runs together.

Whew!

This does concern me a little, as I had previously been using this stanza to “simply” change the default user password to something a little less complicated:


#cloud-config
ssh_pwauth: True
chpasswd:
  list: |
    ubuntu:{{ default_password }}
  expire: False

But now that I look at the documentation, I realise you can also specify that as a pre-hashed value (in which case, you would suffix that default_password item above with |password_hash('sha512')) which makes it all better again!

[1] If you run openstack --os-cloud cloud_a console url show servername gives you a URL to visit that has an HTML5 based VNC-ish client. Note the “cloud_a” and “servername” should be replaced by your clouds.yml reference and the server name or server ID you want to connect to.
[2] Like before, openstack --os-cloud cloud_a console log show servername gives you the output of the boot sequence (e.g. dmesg plus the normal startup commands, and finally, cloud-init). It can be useful. Equally, it’s logs… which means there’s a lot to wade through!

Working with complicated template data UserData in Ansible

JonTheNiceGuy Leave a comment

My new job means I’m currently building a lot of test boxes with Ansible, particularly OpenStack guests. This means I’m trying to script as much as possible without actually … getting my hands dirty with the actual “logging into it and running things” perspective.

This week, I hit a problem standing up a popular firewall vendor’s machine with Ansible, because I was trying to bypass the first-time-wizard… anyway, it wasn’t working, and I couldn’t figure out why. I talked to my colleague [mohclips] and he eventually told me that I needed to use a template, because what I was trying to do was too complicated.

But, damn him, I knew that wasn’t the answer :)

Anyway, I found this comment on a ticket, which lead me to the following… if you’re finding that your userdata: variable in the os_server module of Ansible isn’t working, you might need to wrap it up like this:

userdata: |
  {%- raw -%}#!/bin/bash
  # Kill script if the pipe fails
  set -euf -o pipefail
  # Write everything from this point on to Syslog
  echo " == Set admin credentials == "
  clish -c 'set user admin password-hash {% endraw -%}{{ default_password|password_hash('sha512') }}{%- raw -%}' -s
  {% endraw %}

Note that, if you have a space before your variable, use {% endraw -%} and if you’ve a space after it, use {%- raw %} as the hyphen means “ditch all the spaces before/after this command”.