import more posts from vpotato

This commit is contained in:
John Bowdre 2023-08-15 09:38:41 -05:00
parent de3c5a9331
commit e41457a226
414 changed files with 10179 additions and 2 deletions

View file

@ -0,0 +1,45 @@
---
title: "Accessing a Tanzu Community Edition Kubernetes Cluster from a new device" # Title of the blog post.
date: 2022-02-01T10:58:57-06:00 # Date of post creation.
# lastmod: 2022-02-01T10:58:57-06:00 # Date when last modified
description: "The Tanzu Community Edition documentation does a great job of explaining how to authenticate to a newly-deployed cluster at the tail end of the installation steps, but how do you log in from another system?" # Description used for search engine.
featured: false # Sets if post is a featured post, making appear on the home page side bar.
draft: true # Sets whether to render this page. Draft of true will not be rendered.
toc: false # Controls if a table of contents should be generated for first-level links automatically.
usePageBundles: true
# menu: main
# featureImage: "file.png" # Sets featured image on blog post.
# featureImageAlt: 'Description of image' # Alternative text for featured image.
# featureImageCap: 'This is the featured image.' # Caption (optional).
# thumbnail: "thumbnail.png" # Sets thumbnail image appearing inside card on homepage.
# shareImage: "share.png" # Designate a separate image for social media sharing.
codeLineNumbers: false # Override global value for showing of line numbers within code block.
series: Tips
tags:
- vmware
- kubernetes
- tanzu
comment: true # Disable comment if false.
---
When I [recently set up my Tanzu Community Edition environment](/tanzu-community-edition-k8s-homelab/), I did so from a Linux VM since I knew that my Chromebook Linux environment wouldn't support the `kind` bootstrap cluster used for the deployment. But now I'd like to be able to connect to the cluster directly using the `tanzu` and `kubectl` CLI tools. How do I get the appropriate cluster configuration over to my Chromebook?
The Tanzu CLI actually makes that pretty easy. I just run these commands on my Linux VM to export the `kubeconfig` of my management (`tce-mgmt`) and workload (`tce-work`) clusters to a pair of files:
```shell
tanzu management-cluster kubeconfig get --admin --export-file tce-mgmt-kubeconfig.yaml
tanzu cluster kubeconfig get tce-work --admin --export-file tce-work-kubeconfig.yaml
```
I could then use `scp` to pull the files from the VM into my local Linux environment. I then needed to [install `kubectl`](/tanzu-community-edition-k8s-homelab/#kubectl-binary) and the [`tanzu` CLI](/tanzu-community-edition-k8s-homelab/#tanzu-cli) (making sure to also [enable shell auto-completion](/enable-tanzu-cli-auto-completion-bash-zsh/) along the way!), and I could import the configurations locally:
```shell
tanzu login --kubeconfig tce-mgmt-kubeconfig.yaml --context tce-mgmt-admin@tce-mgmt --name tce-mgmt
✔ successfully logged in to management cluster using the kubeconfig tce-mgmt
tanzu login --kubeconfig tce-work-kubeconfig.yaml --context tce-work-admin@tce-work --name tce-work
✔ successfully logged in to management cluster using the kubeconfig tce-work
```

Binary file not shown.

After

Width:  |  Height:  |  Size: 119 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 134 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 180 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 70 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 214 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 53 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 17 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 41 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 62 KiB

View file

@ -0,0 +1,134 @@
---
series: vRA8
date: "2021-06-01T08:34:30Z"
thumbnail: -Fuvz-GmF.png
usePageBundles: true
tags:
- vmware
- vra
- vro
- javascript
title: Adding VM Notes and Custom Attributes with vRA8
---
*In [past posts](/series/vra8), I started by [creating a basic deployment infrastructure](/vra8-custom-provisioning-part-one) in Cloud Assembly and using tags to group those resources. I then [wrote an integration](/integrating-phpipam-with-vrealize-automation-8) to let vRA8 use phpIPAM for static address assignments. I [implemented a vRO workflow](/vra8-custom-provisioning-part-two) for generating unique VM names which fit an organization's established naming standard, and then [extended the workflow](/vra8-custom-provisioning-part-three) to avoid any naming conflicts in Active Directory and DNS. And, finally, I [created an intelligent provisioning request form in Service Broker](/vra8-custom-provisioning-part-four) to make it easy for users to get the servers they need. That's got the core functionality pretty well sorted, so moving forward I'll be detailing additions that enable new capabilities and enhance the experience.*
In this post, I'll describe how to get certain details from the Service Broker request form and into the VM's properties in vCenter. The obvious application of this is adding descriptive notes so I can remember what purpose a VM serves, but I will also be using [Custom Attributes](https://docs.vmware.com/en/VMware-vSphere/7.0/com.vmware.vsphere.vcenterhost.doc/GUID-73606C4C-763C-4E27-A1DA-032E4C46219D.html) to store the server's Point of Contact information and a record of which ticketing system request resulted in the server's creation.
### New inputs
I'll start this by adding a few new inputs to the cloud template in Cloud Assembly.
![New inputs in Cloud Assembly](F3Wkd3VT.png)
I'm using a basic regex on the `poc_email` field to make sure that the user's input is *probably* a valid email address in the format `[some string]@[some string].[some string]`.
```yaml
inputs:
[...]
description:
type: string
title: Description
description: Server function/purpose
default: Testing and evaluation
poc_name:
type: string
title: Point of Contact Name
default: Jack Shephard
poc_email:
type: string
title: Point of Contact Email
default: jack.shephard@virtuallypotato.com
pattern: '^[^\s@]+@[^\s@]+\.[^\s@]+$'
ticket:
type: string
title: Ticket/Request Number
default: 4815162342
[...]
```
I'll also need to add these to the `resources` section of the template so that they will get passed along with the deployment properties.
![New resource properties](N7YllJkxS.png)
I'm actually going to combine the `poc_name` and `poc_email` fields into a single `poc` string.
```yaml
resources:
Cloud_vSphere_Machine_1:
type: Cloud.vSphere.Machine
properties:
<...>
poc: '${input.poc_name + " (" + input.poc_email + ")"}'
ticket: '${input.ticket}'
description: '${input.description}'
<...>
```
I'll save this as a new version so that the changes will be available in the Service Broker front-end.
![New template version](Z2aKLsLou.png)
### Service Broker custom form
I can then go to Service Broker and drag the new fields onto the Custom Form canvas. (If the new fields don't show up, hit up the Content Sources section of Service Broker, select the content source, and click the "Save and Import" button to sync the changes.) While I'm at it, I set the Description field to display as a text area (encouraging more detailed input), and I also set all the fields on the form to be required.
![Service Broker form](unhgNySSzz.png)
### vRO workflow
Okay, so I've got the information I want to pass on to vCenter. Now I need to whip up a new workflow in vRO that will actually do that (after [telling vRO how to connect to the vCenter](/vra8-custom-provisioning-part-two#interlude-connecting-vro-to-vcenter), of course). I'll want to call this after the VM has been provisioned, so I'll cleverly call the workflow "VM Post-Provisioning".
![Naming the new workflow](X9JhgWx8x.png)
The workflow will have a single input from vRA, `inputProperties` of type `Properties`.
![Workflow input](zHrp6GPcP.png)
The first thing this workflow needs to do is parse `inputProperties (Properties)` to get the name of the VM, and it will then use that information to query vCenter and grab the corresponding VM object. So I'll add a scriptable task item to the workflow canvas and call it `Get VM Object`. It will take `inputProperties (Properties)` as its sole input, and output a new variable called `vm` of type `VC:VirtualMachine`.
![Get VM Object action](5ATk99aPW.png)
The script for this task is fairly straightforward:
```js
// JavaScript: Get VM Object
// Inputs: inputProperties (Properties)
// Outputs: vm (VC:VirtualMachine)
var name = inputProperties.resourceNames[0]
var vms = VcPlugin.getAllVirtualMachines(null, name)
System.log("Found VM object: " + vms[0])
vm = vms[0]
```
I'll add another scriptable task item to the workflow to actually apply the notes to the VM - I'll call it `Set Notes`, and it will take both `vm (VC:VirtualMachine)` and `inputProperties (Properties)` as its inputs.
![Set Notes action](w24V6YVOR.png)
The first part of the script creates a new VM config spec, inserts the description into the spec, and then reconfigures the selected VM with the new spec.
The second part uses a built-in action to set the `Point of Contact` and `Ticket` custom attributes accordingly.
```js
// Javascript: Set Notes
// Inputs: vm (VC:VirtualMachine), inputProperties (Properties)
// Outputs: None
var notes = inputProperties.customProperties.description
var poc = inputProperties.customProperties.poc
var ticket = inputProperties.customProperties.ticket
var spec = new VcVirtualMachineConfigSpec()
spec.annotation = notes
vm.reconfigVM_Task(spec)
System.getModule("com.vmware.library.vc.customattribute").setOrCreateCustomField(vm,"Point of Contact", poc)
System.getModule("com.vmware.library.vc.customattribute").setOrCreateCustomField(vm,"Ticket", ticket)
```
### Extensibility subscription
Now I need to return to Cloud Assembly and create a new extensibility subscription that will call this new workflow at the appropriate time. I'll call it "VM Post-Provisioning" and attach it to the "Compute Post Provision" topic.
![Creating the new subscription](PmhVOWJsUn.png)
And then I'll link it to my new workflow:
![Selecting the workflow](cEbWSOg00.png)
### Testing
And then back to Service Broker to request a VM and see if it works:
![Test request](Lq9DBCK_Y.png)
It worked!
![New VM with notes](-Fuvz-GmF.png)
In the future, I'll be exploring more features that I can add on to this "VM Post-Provisioning" workflow like creating static DNS records as needed.

Binary file not shown.

After

Width:  |  Height:  |  Size: 165 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 157 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 43 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 562 KiB

View file

@ -0,0 +1,98 @@
---
series: Scripts
date: "2021-04-29T08:34:30Z"
usePageBundles: true
thumbnail: 20210723-script.png
tags:
- linux
- shell
- automation
title: Automatic unattended expansion of Linux root LVM volume to fill disk
toc: false
---
While working on my [vRealize Automation 8 project](/series/vra8), I wanted to let users specify how large a VM's system drive should be and have vRA apply that without any further user intervention. For instance, if the template has a 60GB C: drive and the user specifies that they want it to be 80GB, vRA will embiggen the new VM's VMDK to 80GB and then expand the guest file system to fill up the new free space.
I'll get into the details of how that's implemented from the vRA side #soon, but first I needed to come up with simple scripts to extend the guest file system to fill the disk.
This was pretty straight-forward on Windows with a short PowerShell script to grab the appropriate volume and resize it to its full capacity:
```powershell
$Partition = Get-Volume -DriveLetter C | Get-Partition
$Partition | Resize-Partition -Size ($Partition | Get-PartitionSupportedSize).sizeMax
```
It was a bit trickier for Linux systems though. My Linux templates all use LVM to abstract the file systems away from the physical disks, but they may have a different number of physical partitions or different names for the volume groups and logical volumes. So I needed to be able to automagically determine which logical volume was mounted as `/`, which volume group it was a member of, and which partition on which disk is used for that physical volume. I could then expand the physical partition to fill the disk, expand the volume group to fill the now-larger physical volume, grow the logical volume to fill the volume group, and (finally) extend the file system to fill the logical volume.
I found a great script [here](https://github.com/alpacacode/Homebrewn-Scripts/blob/master/linux-scripts/partresize.sh) that helped with most of those operations, but it required the user to specify the physical and logical volumes. I modified it to auto-detect those, and here's what I came up with:
{{% notice info "MBR only" %}}
When I cobbled together this script I was primarily targeting the Enterprise Linux (RHEL, CentOS) systems that I work with in my environment, and those happened to have MBR partition tables. This script would need to be modified a bit to work with GPT partitions like you might find on Ubuntu.
{{% /notice %}}
```shell
#!/bin/bash
# This will attempt to automatically detect the LVM logical volume where / is mounted and then
# expand the underlying physical partition, LVM physical volume, LVM volume group, LVM logical
# volume, and Linux filesystem to consume new free space on the disk.
# Adapted from https://github.com/alpacacode/Homebrewn-Scripts/blob/master/linux-scripts/partresize.sh
extenddisk() {
echo -e "\n+++Current partition layout of $disk:+++"
parted $disk --script unit s print
if [ $logical == 1 ]; then
parted $disk --script rm $ext_partnum
parted $disk --script "mkpart extended ${ext_startsector}s -1s"
parted $disk --script "set $ext_partnum lba off"
parted $disk --script "mkpart logical ext2 ${startsector}s -1s"
else
parted $disk --script rm $partnum
parted $disk --script "mkpart primary ext2 ${startsector}s -1s"
fi
parted $disk --script set $partnum lvm on
echo -e "\n\n+++New partition layout of $disk:+++"
parted $disk --script unit s print
partx -v -a $disk
pvresize $pvname
lvextend --extents +100%FREE --resize $lvpath
echo -e "\n+++New root partition size:+++"
df -h / | grep -v Filesystem
}
export LVM_SUPPRESS_FD_WARNINGS=1
mountpoint=$(df --output=source / | grep -v Filesystem) # /dev/mapper/centos-root
lvdisplay $mountpoint > /dev/null
if [ $? != 0 ]; then
echo "Error: $mountpoint does not look like a LVM logical volume. Aborting."
exit 1
fi
echo -e "\n+++Current root partition size:+++"
df -h / | grep -v Filesystem
lvname=$(lvs --noheadings $mountpoint | awk '{print($1)}') # root
vgname=$(lvs --noheadings $mountpoint | awk '{print($2)}') # centos
lvpath="/dev/${vgname}/${lvname}" # /dev/centos/root
pvname=$(pvs | grep $vgname | tail -n1 | awk '{print($1)}') # /dev/sda2
disk=$(echo $pvname | rev | cut -c 2- | rev) # /dev/sda
diskshort=$(echo $disk | grep -Po '[^\/]+$') # sda
partnum=$(echo $pvname | grep -Po '\d$') # 2
startsector=$(fdisk -u -l $disk | grep $pvname | awk '{print $2}') # 2099200
layout=$(parted $disk --script unit s print) # Model: VMware Virtual disk (scsi) Disk /dev/sda: 83886080s Sector size (logical/physical): 512B/512B Partition Table: msdos Disk Flags: Number Start End Size Type File system Flags 1 2048s 2099199s 2097152s primary xfs boot 2 2099200s 62914559s 60815360s primary lvm
if grep -Pq "^\s$partnum\s+.+?logical.+$" <<< "$layout"; then
logical=1
ext_partnum=$(parted $disk --script unit s print | grep extended | grep -Po '^\s\d\s' | tr -d ' ')
ext_startsector=$(parted $disk --script unit s print | grep extended | awk '{print $2}' | tr -d 's')
else
logical=0
fi
parted $disk --script unit s print | if ! grep -Pq "^\s$partnum\s+.+?[^,]+?lvm\s*$"; then
echo -e "Error: $pvname seems to have some flags other than 'lvm' set."
exit 1
fi
if ! (fdisk -u -l $disk | grep $disk | tail -1 | grep $pvname | grep -q "Linux LVM"); then
echo -e "Error: $pvname is not the last LVM volume on disk $disk."
exit 1
fi
ls /sys/class/scsi_device/*/device/rescan | while read path; do echo 1 > $path; done
ls /sys/class/scsi_host/host*/scan | while read path; do echo "- - -" > $path; done
extenddisk
```
And it works beautifully within my environment. Hopefully it'll work for yours too in case you have a similar need!

Binary file not shown.

After

Width:  |  Height:  |  Size: 828 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 648 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 308 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.3 MiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.1 MiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 238 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.2 MiB

View file

@ -0,0 +1,620 @@
---
title: "Bulk Import vSphere dvPortGroups to phpIPAM" # Title of the blog post.
date: 2022-02-04 # Date of post creation.
# lastmod: 2022-01-21T15:24:00-06:00 # Date when last modified
description: "I wrote a Python script to interface with the phpIPAM API and import a large number of networks exported from vSphere for IP management." # Description used for search engine.
featured: false # Sets if post is a featured post, making appear on the home page side bar.
draft: false # Sets whether to render this page. Draft of true will not be rendered.
toc: true # Controls if a table of contents should be generated for first-level links automatically.
usePageBundles: true
# menu: main
# featureImage: "file.png" # Sets featured image on blog post.
# featureImageAlt: 'Description of image' # Alternative text for featured image.
# featureImageCap: 'This is the featured image.' # Caption (optional).
thumbnail: "code.png" # Sets thumbnail image appearing inside card on homepage.
# shareImage: "share.png" # Designate a separate image for social media sharing.
codeLineNumbers: false # Override global value for showing of line numbers within code block.
series: Scripts
tags:
- vmware
- powercli
- python
- api
- phpipam
comment: true # Disable comment if false.
---
I [recently wrote](/tanzu-community-edition-k8s-homelab/#a-real-workload---phpipam) about getting started with VMware's [Tanzu Community Edition](https://tanzucommunityedition.io/) and deploying [phpIPAM](https://phpipam.net/) as my first real-world Kubernetes workload. Well I've spent much of my time since then working on a script which would help to populate my phpIPAM instance with a list of networks to monitor.
### Planning and Exporting
The first step in making this work was to figure out which networks I wanted to import. We've got hundreds of different networks in use across our production vSphere environments. I focused only on those which are portgroups on distributed virtual switches since those configurations are pretty standardized (being vCenter constructs instead of configured on individual hosts). These dvPortGroups bear a naming standard which conveys all sorts of useful information, and it's easy and safe to rename any dvPortGroups which _don't_ fit the standard (unlike renaming portgroups on a standard virtual switch).
The standard naming convention is `[Site/Description] [Network Address]{/[Mask]}`. So the networks (across two virtual datacenters and two dvSwitches) look something like this:
![Production dvPortGroups approximated in my testing lab environment](dvportgroups.png)
Some networks have masks in the name, some don't; and some use an underscore (`_`) rather than a slash (`/`) to separate the network from the mask . Most networks correctly include the network address with a `0` in the last octet, but some use an `x` instead. And the VLANs associated with the networks have a varying number of digits. Consistency can be difficult so these are all things that I had to keep in mind as I worked on a solution which would make a true best effort at importing all of these.
As long as the dvPortGroup names stick to this format I can parse the name to come up with a description as well as the IP space of the network. The dvPortGroup also carries information about the associated VLAN, which is useful information to have. And I can easily export this information with a simple PowerCLI query:
```powershell
PS /home/john> get-vdportgroup | select Name, VlanConfiguration
Name VlanConfiguration
---- -----------------
MGT-Home 192.168.1.0
MGT-Servers 172.16.10.0 VLAN 1610
BOW-Servers 172.16.20.0 VLAN 1620
BOW-Servers 172.16.30.0 VLAN 1630
BOW-Servers 172.16.40.0 VLAN 1640
DRE-Servers 172.16.50.0 VLAN 1650
DRE-Servers 172.16.60.x VLAN 1660
VPOT8-Mgmt 172.20.10.0/27 VLAN 20
VPOT8-Servers 172.20.10.32/27 VLAN 30
VPOT8-Servers 172.20.10.64_26 VLAN 40
```
In my [homelab](/vmware-home-lab-on-intel-nuc-9/), I only have a single vCenter. In production, we've got a handful of vCenters, and each manages the hosts in a given region. So I can use information about which vCenter hosts a dvPortGroup to figure out which region a network is in. When I import this data into phpIPAM, I can use the vCenter name to assign [remote scan agents](https://github.com/jbowdre/phpipam-agent-docker) to networks based on the region that they're in. I can also grab information about which virtual datacenter a dvPortGroup lives in, which I'll use for grouping networks into sites or sections.
The vCenter can be found in the `Uid` property returned by `get-vdportgroup`:
```powershell
PS /home/john> get-vdportgroup | select Name, VlanConfiguration, Datacenter, Uid
Name VlanConfiguration Datacenter Uid
---- ----------------- ---------- ---
MGT-Home 192.168.1.0 Lab /VIServer=lab\john@vcsa.lab.bowdre.net:443/DistributedPortgroup=DistributedVirtualPortgroup-dvportgroup-27015/
MGT-Servers 172.16.10.0 VLAN 1610 Lab /VIServer=lab\john@vcsa.lab.bowdre.net:443/DistributedPortgroup=DistributedVirtualPortgroup-dvportgroup-27017/
BOW-Servers 172.16.20.0 VLAN 1620 Lab /VIServer=lab\john@vcsa.lab.bowdre.net:443/DistributedPortgroup=DistributedVirtualPortgroup-dvportgroup-28010/
BOW-Servers 172.16.30.0 VLAN 1630 Lab /VIServer=lab\john@vcsa.lab.bowdre.net:443/DistributedPortgroup=DistributedVirtualPortgroup-dvportgroup-28011/
BOW-Servers 172.16.40.0 VLAN 1640 Lab /VIServer=lab\john@vcsa.lab.bowdre.net:443/DistributedPortgroup=DistributedVirtualPortgroup-dvportgroup-28012/
DRE-Servers 172.16.50.0 VLAN 1650 Lab /VIServer=lab\john@vcsa.lab.bowdre.net:443/DistributedPortgroup=DistributedVirtualPortgroup-dvportgroup-28013/
DRE-Servers 172.16.60.x VLAN 1660 Lab /VIServer=lab\john@vcsa.lab.bowdre.net:443/DistributedPortgroup=DistributedVirtualPortgroup-dvportgroup-28014/
VPOT8-Mgmt 172.20.10.0/… VLAN 20 Other Lab /VIServer=lab\john@vcsa.lab.bowdre.net:443/DistributedPortgroup=DistributedVirtualPortgroup-dvportgroup-35018/
VPOT8-Servers 172.20.10… VLAN 30 Other Lab /VIServer=lab\john@vcsa.lab.bowdre.net:443/DistributedPortgroup=DistributedVirtualPortgroup-dvportgroup-35019/
VPOT8-Servers 172.20.10… VLAN 40 Other Lab /VIServer=lab\john@vcsa.lab.bowdre.net:443/DistributedPortgroup=DistributedVirtualPortgroup-dvportgroup-35020/
```
It's not pretty, but it'll do the trick. All that's left is to export this data into a handy-dandy CSV-formatted file that I can easily parse for import:
```powershell
get-vdportgroup | select Name, VlanConfiguration, Datacenter, Uid | export-csv -NoTypeInformation ./networks.csv
```
![My networks.csv export, including the networks which don't match the naming criteria and will be skipped by the import process.](networks.csv.png)
### Setting up phpIPAM
After [deploying a fresh phpIPAM instance on my Tanzu Community Edition Kubernetes cluster](/tanzu-community-edition-k8s-homelab/#a-real-workload---phpipam), there are a few additional steps needed to enable API access. To start, I log in to my phpIPAM instance and navigate to the **Administration > Server Management > phpIPAM Settings** page, where I enabled both the *Prettify links* and *API* feature settings - making sure to hit the **Save** button at the bottom of the page once I do so.
![Enabling the API](server_settings.png)
Then I need to head to the **User Management** page to create a new user that will be used to authenticate against the API:
![New user creation](new_user.png)
And finally, I head to the **API** section to create a new API key with Read/Write permissions:
![API key creation](api_user.png)
I'm also going to head in to **Administration > IP Related Management > Sections** and delete the default sample sections so that the inventory will be nice and empty:
![We don't need no stinkin' sections!](empty_sections.png)
### Script time
Well that's enough prep work; now it's time for the Python3 [script](https://github.com/jbowdre/misc-scripts/blob/main/Python/phpipam-bulk-import.py):
```python
# The latest version of this script can be found on Github:
# https://github.com/jbowdre/misc-scripts/blob/main/Python/phpipam-bulk-import.py
import requests
from collections import namedtuple
check_cert = True
created = 0
remote_agent = False
name_to_id = namedtuple('name_to_id', ['name', 'id'])
## for testing only:
# from requests.packages.urllib3.exceptions import InsecureRequestWarning
# requests.packages.urllib3.disable_warnings(InsecureRequestWarning)
# check_cert = False
## Makes sure input fields aren't blank.
def validate_input_is_not_empty(field, prompt):
while True:
user_input = input(f'\n{prompt}:\n')
if len(user_input) == 0:
print(f'[ERROR] {field} cannot be empty!')
continue
else:
return user_input
## Takes in a list of dictionary items, extracts all the unique values for a given key,
# and returns a sorted list of those.
def get_sorted_list_of_unique_values(key, list_of_dict):
valueSet = set(sub[key] for sub in list_of_dict)
valueList = list(valueSet)
valueList.sort()
return valueList
## Match names and IDs
def get_id_from_sets(name, sets):
return [item.id for item in sets if name == item.name][0]
## Authenticate to phpIPAM endpoint and return an auth token
def auth_session(uri, auth):
print(f'Authenticating to {uri}...')
try:
req = requests.post(f'{uri}/user/', auth=auth, verify=check_cert)
except:
raise requests.exceptions.RequestException
if req.status_code != 200:
print(f'[ERROR] Authentication failure: {req.json()}')
raise requests.exceptions.RequestException
token = {"token": req.json()['data']['token']}
print('\n[AUTH_SUCCESS] Authenticated successfully!')
return token
## Find or create a remote scan agent for each region (vcenter)
def get_agent_sets(uri, token, regions):
agent_sets = []
def create_agent_set(uri, token, name):
import secrets
# generate a random secret to be used for identifying this agent
payload = {
'name': name,
'type': 'mysql',
'code': secrets.base64.urlsafe_b64encode(secrets.token_bytes(24)).decode("utf-8"),
'description': f'Remote scan agent for region {name}'
}
req = requests.post(f'{uri}/tools/scanagents/', data=payload, headers=token, verify=check_cert)
id = req.json()['id']
agent_set = name_to_id(name, id)
print(f'[AGENT_CREATE] {name} created.')
return agent_set
for region in regions:
name = regions[region]['name']
req = requests.get(f'{uri}/tools/scanagents/?filter_by=name&filter_value={name}', headers=token, verify=check_cert)
if req.status_code == 200:
id = req.json()['data'][0]['id']
agent_set = name_to_id(name, id)
else:
agent_set = create_agent_set(uri, token, name)
agent_sets.append(agent_set)
return agent_sets
## Find or create a section for each virtual datacenter
def get_section(uri, token, section, parentSectionId):
def create_section(uri, token, section, parentSectionId):
payload = {
'name': section,
'masterSection': parentSectionId,
'permissions': '{"2":"2"}',
'showVLAN': '1'
}
req = requests.post(f'{uri}/sections/', data=payload, headers=token, verify=check_cert)
id = req.json()['id']
print(f'[SECTION_CREATE] Section {section} created.')
return id
req = requests.get(f'{uri}/sections/{section}/', headers=token, verify=check_cert)
if req.status_code == 200:
id = req.json()['data']['id']
else:
id = create_section(uri, token, section, parentSectionId)
return id
## Find or create VLANs
def get_vlan_sets(uri, token, vlans):
vlan_sets = []
def create_vlan_set(uri, token, vlan):
payload = {
'name': f'VLAN {vlan}',
'number': vlan
}
req = requests.post(f'{uri}/vlan/', data=payload, headers=token, verify=check_cert)
id = req.json()['id']
vlan_set = name_to_id(vlan, id)
print(f'[VLAN_CREATE] VLAN {vlan} created.')
return vlan_set
for vlan in vlans:
if vlan != 0:
req = requests.get(f'{uri}/vlan/?filter_by=number&filter_value={vlan}', headers=token, verify=check_cert)
if req.status_code == 200:
id = req.json()['data'][0]['vlanId']
vlan_set = name_to_id(vlan, id)
else:
vlan_set = create_vlan_set(uri, token, vlan)
vlan_sets.append(vlan_set)
return vlan_sets
## Find or create nameserver configurations for each region
def get_nameserver_sets(uri, token, regions):
nameserver_sets = []
def create_nameserver_set(uri, token, name, nameservers):
payload = {
'name': name,
'namesrv1': nameservers,
'description': f'Nameserver created for region {name}'
}
req = requests.post(f'{uri}/tools/nameservers/', data=payload, headers=token, verify=check_cert)
id = req.json()['id']
nameserver_set = name_to_id(name, id)
print(f'[NAMESERVER_CREATE] Nameserver {name} created.')
return nameserver_set
for region in regions:
name = regions[region]['name']
req = requests.get(f'{uri}/tools/nameservers/?filter_by=name&filter_value={name}', headers=token, verify=check_cert)
if req.status_code == 200:
id = req.json()['data'][0]['id']
nameserver_set = name_to_id(name, id)
else:
nameserver_set = create_nameserver_set(uri, token, name, regions[region]['nameservers'])
nameserver_sets.append(nameserver_set)
return nameserver_sets
## Find or create subnet for each dvPortGroup
def create_subnet(uri, token, network):
def update_nameserver_permissions(uri, token, network):
nameserverId = network['nameserverId']
sectionId = network['sectionId']
req = requests.get(f'{uri}/tools/nameservers/{nameserverId}/', headers=token, verify=check_cert)
permissions = req.json()['data']['permissions']
permissions = str(permissions).split(';')
if not sectionId in permissions:
permissions.append(sectionId)
if 'None' in permissions:
permissions.remove('None')
permissions = ';'.join(permissions)
payload = {
'permissions': permissions
}
req = requests.patch(f'{uri}/tools/nameservers/{nameserverId}/', data=payload, headers=token, verify=check_cert)
payload = {
'subnet': network['subnet'],
'mask': network['mask'],
'description': network['name'],
'sectionId': network['sectionId'],
'scanAgent': network['agentId'],
'nameserverId': network['nameserverId'],
'vlanId': network['vlanId'],
'pingSubnet': '1',
'discoverSubnet': '1',
'resolveDNS': '1',
'DNSrecords': '1'
}
req = requests.post(f'{uri}/subnets/', data=payload, headers=token, verify=check_cert)
if req.status_code == 201:
network['subnetId'] = req.json()['id']
update_nameserver_permissions(uri, token, network)
print(f"[SUBNET_CREATE] Created subnet {req.json()['data']}")
global created
created += 1
elif req.status_code == 409:
print(f"[SUBNET_EXISTS] Subnet {network['subnet']}/{network['mask']} already exists.")
else:
print(f"[ERROR] Problem creating subnet {network['name']}: {req.json()}")
## Import list of networks from the specified CSV file
def import_networks(filepath):
print(f'Importing networks from {filepath}...')
import csv
import re
ipPattern = re.compile('\d{1,3}\.\d{1,3}\.\d{1,3}\.[0-9xX]{1,3}')
networks = []
with open(filepath) as csv_file:
reader = csv.DictReader(csv_file)
line_count = 0
for row in reader:
network = {}
if line_count > 0:
if(re.search(ipPattern, row['Name'])):
network['subnet'] = re.findall(ipPattern, row['Name'])[0]
if network['subnet'].split('.')[-1].lower() == 'x':
network['subnet'] = network['subnet'].lower().replace('x', '0')
network['name'] = row['Name']
if '/' in row['Name'][-3]:
network['mask'] = row['Name'].split('/')[-1]
elif '_' in row['Name'][-3]:
network['mask'] = row['Name'].split('_')[-1]
else:
network['mask'] = '24'
network['section'] = row['Datacenter']
try:
network['vlan'] = int(row['VlanConfiguration'].split('VLAN ')[1])
except:
network['vlan'] = 0
network['vcenter'] = f"{(row['Uid'].split('@'))[1].split(':')[0].split('.')[0]}"
networks.append(network)
line_count += 1
print(f'Processed {line_count} lines and found:')
return networks
def main():
import socket
import getpass
import argparse
from pathlib import Path
parser = argparse.ArgumentParser()
parser.add_argument("filepath", type=Path)
# Accept CSV file as an argument to the script or prompt for input if necessary
try:
p = parser.parse_args()
filepath = p.filepath
except:
# make sure filepath is a path to an actual file
print("""\n\n
This script helps to add vSphere networks to phpIPAM for IP address management. It is expected
that the vSphere networks are configured as portgroups on distributed virtual switches and
named like '[Description] [Subnet IP]{/[mask]}' (ex: 'LAB-Servers 192.168.1.0'). The following PowerCLI
command can be used to export the networks from vSphere:
Get-VDPortgroup | Select Name, Datacenter, VlanConfiguration, Uid | Export-Csv -NoTypeInformation ./networks.csv
Subnets added to phpIPAM will be automatically configured for monitoring either using the built-in
scan agent (default) or a new remote scan agent for each vCenter.
""")
while True:
filepath = Path(validate_input_is_not_empty('Filepath', 'Path to CSV-formatted export from vCenter'))
if filepath.exists():
break
else:
print(f'[ERROR] Unable to find file at {filepath.name}.')
continue
# get collection of networks to import
networks = import_networks(filepath)
networkNames = get_sorted_list_of_unique_values('name', networks)
print(f'\n- {len(networkNames)} networks:\n\t{networkNames}')
vcenters = get_sorted_list_of_unique_values('vcenter', networks)
print(f'\n- {len(vcenters)} vCenter servers:\n\t{vcenters}')
vlans = get_sorted_list_of_unique_values('vlan', networks)
print(f'\n- {len(vlans)} VLANs:\n\t{vlans}')
sections = get_sorted_list_of_unique_values('section', networks)
print(f'\n- {len(sections)} Datacenters:\n\t{sections}')
regions = {}
for vcenter in vcenters:
nameservers = None
name = validate_input_is_not_empty('Region Name', f'Region name for vCenter {vcenter}')
for region in regions:
if name in regions[region]['name']:
nameservers = regions[region]['nameservers']
if not nameservers:
nameservers = validate_input_is_not_empty('Nameserver IPs', f"Comma-separated list of nameserver IPs in {name}")
nameservers = nameservers.replace(',',';').replace(' ','')
regions[vcenter] = {'name': name, 'nameservers': nameservers}
# make sure hostname resolves
while True:
hostname = input('\nFully-qualified domain name of the phpIPAM host:\n')
if len(hostname) == 0:
print('[ERROR] Hostname cannot be empty.')
continue
try:
test = socket.gethostbyname(hostname)
except:
print(f'[ERROR] Unable to resolve {hostname}.')
continue
else:
del test
break
username = validate_input_is_not_empty('Username', f'Username with read/write access to {hostname}')
password = getpass.getpass(f'Password for {username}:\n')
apiAppId = validate_input_is_not_empty('App ID', f'App ID for API key (from https://{hostname}/administration/api/)')
agent = input('\nUse per-region remote scan agents instead of a single local scanner? (y/N):\n')
try:
if agent.lower()[0] == 'y':
global remote_agent
remote_agent = True
except:
pass
proceed = input(f'\n\nProceed with importing {len(networkNames)} networks to {hostname}? (y/N):\n')
try:
if proceed.lower()[0] == 'y':
pass
else:
import sys
sys.exit("Operation aborted.")
except:
import sys
sys.exit("Operation aborted.")
del proceed
# assemble variables
uri = f'https://{hostname}/api/{apiAppId}'
auth = (username, password)
# auth to phpIPAM
token = auth_session(uri, auth)
# create nameserver entries
nameserver_sets = get_nameserver_sets(uri, token, regions)
vlan_sets = get_vlan_sets(uri, token, vlans)
if remote_agent:
agent_sets = get_agent_sets(uri, token, regions)
# create the networks
for network in networks:
network['region'] = regions[network['vcenter']]['name']
network['regionId'] = get_section(uri, token, network['region'], None)
network['nameserverId'] = get_id_from_sets(network['region'], nameserver_sets)
network['sectionId'] = get_section(uri, token, network['section'], network['regionId'])
if network['vlan'] == 0:
network['vlanId'] = None
else:
network['vlanId'] = get_id_from_sets(network['vlan'], vlan_sets)
if remote_agent:
network['agentId'] = get_id_from_sets(network['region'], agent_sets)
else:
network['agentId'] = '1'
create_subnet(uri, token, network)
print(f'\n[FINISH] Created {created} of {len(networks)} networks.')
if __name__ == "__main__":
main()
```
I'll run it and provide the path to the network export CSV file:
```bash
python3 phpipam-bulk-import.py ~/networks.csv
```
The script will print out a little descriptive bit about what sort of networks it's going to try to import and then will straight away start processing the file to identify the networks, vCenters, VLANs, and datacenters which will be imported:
```
Importing networks from /home/john/networks.csv...
Processed 17 lines and found:
- 10 networks:
['BOW-Servers 172.16.20.0', 'BOW-Servers 172.16.30.0', 'BOW-Servers 172.16.40.0', 'DRE-Servers 172.16.50.0', 'DRE-Servers 172.16.60.x', 'MGT-Home 192.168.1.0', 'MGT-Servers 172.16.10.0', 'VPOT8-Mgmt 172.20.10.0/27', 'VPOT8-Servers 172.20.10.32/27', 'VPOT8-Servers 172.20.10.64_26']
- 1 vCenter servers:
['vcsa']
- 10 VLANs:
[0, 20, 30, 40, 1610, 1620, 1630, 1640, 1650, 1660]
- 2 Datacenters:
['Lab', 'Other Lab']
```
It then starts prompting for the additional details which will be needed:
```
Region name for vCenter vcsa:
Labby
Comma-separated list of nameserver IPs in Lab vCenter:
192.168.1.5
Fully-qualified domain name of the phpIPAM host:
ipam-k8s.lab.bowdre.net
Username with read/write access to ipam-k8s.lab.bowdre.net:
api-user
Password for api-user:
App ID for API key (from https://ipam-k8s.lab.bowdre.net/administration/api/):
api-user
Use per-region remote scan agents instead of a single local scanner? (y/N):
y
```
Up to this point, the script has only been processing data locally, getting things ready for talking to the phpIPAM API. But now, it prompts to confirm that we actually want to do the thing (yes please) and then gets to work:
```
Proceed with importing 10 networks to ipam-k8s.lab.bowdre.net? (y/N):
y
Authenticating to https://ipam-k8s.lab.bowdre.net/api/api-user...
[AUTH_SUCCESS] Authenticated successfully!
[VLAN_CREATE] VLAN 20 created.
[VLAN_CREATE] VLAN 30 created.
[VLAN_CREATE] VLAN 40 created.
[VLAN_CREATE] VLAN 1610 created.
[VLAN_CREATE] VLAN 1620 created.
[VLAN_CREATE] VLAN 1630 created.
[VLAN_CREATE] VLAN 1640 created.
[VLAN_CREATE] VLAN 1650 created.
[VLAN_CREATE] VLAN 1660 created.
[SECTION_CREATE] Section Labby created.
[SECTION_CREATE] Section Lab created.
[SUBNET_CREATE] Created subnet 192.168.1.0/24
[SUBNET_CREATE] Created subnet 172.16.10.0/24
[SUBNET_CREATE] Created subnet 172.16.20.0/24
[SUBNET_CREATE] Created subnet 172.16.30.0/24
[SUBNET_CREATE] Created subnet 172.16.40.0/24
[SUBNET_CREATE] Created subnet 172.16.50.0/24
[SUBNET_CREATE] Created subnet 172.16.60.0/24
[SECTION_CREATE] Section Other Lab created.
[SUBNET_CREATE] Created subnet 172.20.10.0/27
[SUBNET_CREATE] Created subnet 172.20.10.32/27
[SUBNET_CREATE] Created subnet 172.20.10.64/26
[FINISH] Created 10 of 10 networks.
```
Success! Now I can log in to my phpIPAM instance and check out my newly-imported subnets:
![New subnets!](created_subnets.png)
Even the one with the weird name formatting was parsed and imported correctly:
![Subnet details](subnet_detail.png)
So now phpIPAM knows about the vSphere networks I care about, and it can keep track of which vLAN and nameservers go with which networks. Great! But it still isn't scanning or monitoring those networks, even though I told the script that I wanted to use a remote scan agent. And I can check in the **Administration > Server management > Scan agents** section of the phpIPAM interface to see my newly-created agent configuration.
![New agent config](agent_config.png)
... but I haven't actually *deployed* an agent yet. I'll do that by following the same basic steps [described here](/tanzu-community-edition-k8s-homelab/#phpipam-agent) to spin up my `phpipam-agent` on Kubernetes, and I'll plug in that automagically-generated code for the `IPAM_AGENT_KEY` environment variable:
```yaml
---
apiVersion: apps/v1
kind: Deployment
metadata:
name: phpipam-agent
spec:
selector:
matchLabels:
app: phpipam-agent
replicas: 1
template:
metadata:
labels:
app: phpipam-agent
spec:
containers:
- name: phpipam-agent
image: ghcr.io/jbowdre/phpipam-agent:latest
env:
- name: IPAM_DATABASE_HOST
value: "ipam-k8s.lab.bowdre.net"
- name: IPAM_DATABASE_NAME
value: "phpipam"
- name: IPAM_DATABASE_USER
value: "phpipam"
- name: IPAM_DATABASE_PASS
value: "VMware1!"
- name: IPAM_DATABASE_PORT
value: "3306"
- name: IPAM_AGENT_KEY
value: "CxtRbR81r1ojVL2epG90JaShxIUBl0bT"
- name: IPAM_SCAN_INTERVAL
value: "15m"
- name: IPAM_RESET_AUTODISCOVER
value: "false"
- name: IPAM_REMOVE_DHCP
value: "false"
- name: TZ
value: "UTC"
```
I kick it off with a `kubectl apply` command and check back a few minutes later (after the 15-minute interval defined in the above YAML) to see that it worked, the remote agent scanned like it was supposed to and is reporting IP status back to the phpIPAM database server:
![Newly-discovered IPs](discovered_ips.png)
I think I've got some more tweaks to do with this environment (why isn't phpIPAM resolving hostnames despite the correct DNS servers getting configured?) but this at least demonstrates a successful proof-of-concept import thanks to my Python script. Sure, I only imported 10 networks here, but I feel like I'm ready to process the several hundred which are available in our production environment now.
And who knows, maybe this script will come in handy for someone else. Until next time!

Binary file not shown.

After

Width:  |  Height:  |  Size: 416 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 727 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 247 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 1.6 MiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 889 KiB

View file

@ -2,6 +2,7 @@
series: Tips
date: "2020-12-23T08:34:30Z"
thumbnail: -lp1-DGiM.png
usePageBundles: true
tags:
- chromeos
title: Burn an ISO to USB with the Chromebook Recovery Utility
@ -9,14 +10,14 @@ toc: false
featured: true
---
There are a number of fantastic Windows applications for creating bootable USB drives from ISO images - but those don't work on a Chromebook. Fortunately there's an easily-available tool which will do the trick: Google's own [Chromebook Recovery Utility](https://chrome.google.com/webstore/detail/chromebook-recovery-utili/pocpnlppkickgojjlmhdmidojbmbodfm) app.
There are a number of fantastic Windows applications for creating bootable USB drives from ISO images - but those don't work on a Chromebook. Fortunately there's an easily-available tool which will do the trick: Google's own [Chromebook Recovery Utility](https://chrome.google.com/webstore/detail/chromebook-recovery-utili/pocpnlppkickgojjlmhdmidojbmbodfm) app.
Normally that tool is used to creating bootable media to [reinstall Chrome OS on a broken Chromebook](https://support.google.com/chromebook/answer/1080595) (hence the name) but it also has the capability to write other arbitrary images as well. So if you find yourself needing to create a USB drive for installing ESXi on a computer in your [home lab](https://twitter.com/johndotbowdre/status/1341767090945077248) (more on that soon!) here's what you'll need to do:
1. Install the [Chromebook Recovery Utility](https://chrome.google.com/webstore/detail/chromebook-recovery-utili/pocpnlppkickgojjlmhdmidojbmbodfm).
2. Download the ISO you intend to use.
3. Rename the file to append `.bin` on the end, after the `.iso` bit:
![My renamed ISO for installing ESXi](uoTjgtbN1.png)
![My renamed ISO for installing ESXi](uoTjgtbN1.png)
4. Plug in the USB drive you're going to sacrifice for this effort - remember that ALL data on the drive will be erased.
5. Open the recovery utility, click on the gear icon at the top right, and select the *Use local image* option:
![The CRU menu](vdTpW9t7Q.png)

Binary file not shown.

After

Width:  |  Height:  |  Size: 81 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 242 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 147 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 144 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 54 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 26 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 234 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 124 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 121 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 89 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 219 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 318 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 91 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 116 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 118 KiB

View file

@ -0,0 +1,421 @@
---
series: vRA8
date: "2021-08-13T00:00:00Z"
lastmod: "2022-01-18"
usePageBundles: true
thumbnail: 20210813_workflow_success.png
tags:
- vmware
- vra
- vro
- javascript
- powershell
- automation
title: Creating static records in Microsoft DNS from vRealize Automation
---
One of the requirements for my vRA deployments is the ability to automatically create a static `A` records for non-domain-joined systems so that users can connect without needing to know the IP address. The organization uses Microsoft DNS servers to provide resolution on the internal domain. At first glance, this shouldn't be too much of a problem: vRealize Orchestrator 8.x can run PowerShell scripts, and PowerShell can use the [`Add-DnsServerResourceRecord` cmdlet](https://docs.microsoft.com/en-us/powershell/module/dnsserver/add-dnsserverresourcerecord?view=windowsserver2019-ps) to create the needed records.
Not so fast, though. That cmdlet is provided through the [Remote Server Administration Tools](https://docs.microsoft.com/en-us/troubleshoot/windows-server/system-management-components/remote-server-administration-tools) package so it won't be available within the limited PowerShell environment inside of vRO. A workaround might be to add a Windows machine to vRO as a remote PowerShell host, but then you run into [issues of credential hopping](https://communities.vmware.com/t5/vRealize-Orchestrator/unable-to-run-get-DnsServerResourceRecord-via-vRO-Powershell-wf/m-p/2286685).
I eventually came across [this blog post](https://www.virtualnebula.com/blog/2017/7/14/microsoft-ad-dns-integration-over-ssh) which described adding a Windows machine as a remote *SSH* host instead. I'll deviate a bit from the described configuration, but that post did at least get me pointed in the right direction. This approach would get around the complicated authentication-tunneling business while still being pretty easy to set up. So let's go!
### Preparing the SSH host
I deployed a Windows Server 2019 Core VM to use as my SSH host, and I joined it to my AD domain as `win02.lab.bowdre.net`. Once that's taken care of, I need to install the RSAT DNS tools so that I can use the `Add-DnsServerResourceRecord` and associated cmdlets. I can do that through PowerShell like so:
```powershell
# Install RSAT DNS tools
Add-WindowsCapability -online -name Rsat.Dns.Tools~~~~0.0.1.0
```
Instead of using a third-party SSH server, I'll use the OpenSSH Server that's already available in Windows 10 (1809+) and Server 2019:
```powershell
# Install OpenSSH Server
Add-WindowsCapability -Online -Name OpenSSH.Server~~~~0.0.1.0
```
I'll also want to set it so that the default shell upon SSH login is PowerShell (rather than the standard Command Prompt) so that I can have easy access to those DNS cmdlets:
```powershell
# Set PowerShell as the default Shell (for access to DNS cmdlets)
New-ItemProperty -Path "HKLM:\SOFTWARE\OpenSSH" -Name DefaultShell -Value "C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe" -PropertyType String -Force
```
I'll be using my `lab\vra` service account for managing DNS. I've already given it the appropriate rights on the DNS server, but I'll also add it to the Administrators group on my SSH host:
```powershell
# Add the service account as a local administrator
Add-LocalGroupMember -Group Administrators -Member "lab\vra"
```
And I'll modify the OpenSSH configuration so that only members of that Administrators group are permitted to log into the server via SSH:
```powershell
# Restrict SSH access to members in the local Administrators group
(Get-Content "C:\ProgramData\ssh\sshd_config") -Replace "# Authentication:", "$&`nAllowGroups Administrators" | Set-Content "C:\ProgramData\ssh\sshd_config"
```
Finally, I'll start the `sshd` service and set it to start up automatically:
```powershell
# Start service and set it to automatic
Set-Service -Name sshd -StartupType Automatic -Status Running
```
#### A quick test
At this point, I can log in to the server via SSH and confirm that I can create and delete records in my DNS zone:
```powershell
$ ssh vra@win02.lab.bowdre.net
vra@win02.lab.bowdre.net's password:
Windows PowerShell
Copyright (C) Microsoft Corporation. All rights reserved.
PS C:\Users\vra> Add-DnsServerResourceRecordA -ComputerName win01.lab.bowdre.net -Name testy -ZoneName lab.bowdre.net -AllowUpdateAny -IPv4Address 172.16.99.99
PS C:\Users\vra> nslookup testy
Server: win01.lab.bowdre.net
Address: 192.168.1.5
Name: testy.lab.bowdre.net
Address: 172.16.99.99
PS C:\Users\vra> Remove-DnsServerResourceRecord -ComputerName win01.lab.bowdre.net -Name testy -ZoneName lab.bowdre.net -RRType A -Force
PS C:\Users\vra> nslookup testy
Server: win01.lab.bowdre.net
Address: 192.168.1.5
*** win01.lab.bowdre.net can't find testy: Non-existent domain
```
Cool! Now I just need to do that same thing, but from vRealize Orchestrator. First, though, I'll update the template so the requester can choose whether or not a static record will get created.
### Template changes
#### Cloud Template
Similar to the template changes I made for [optionally joining deployed servers to the Active Directory domain](/joining-vms-to-active-directory-in-site-specific-ous-with-vra8#cloud-template), I'll just be adding a simple boolean checkbox to the `inputs` section of the template in Cloud Assembly:
```yaml
formatVersion: 1
inputs:
[...]
staticDns:
title: Create static DNS record
type: boolean
default: false
[...]
```
*Unlike* the AD piece, in the `resources` section I'll just bind a custom property called `staticDns` to the input with the same name:
```yaml
resources:
Cloud_vSphere_Machine_1:
type: Cloud.vSphere.Machine
properties:
[...]
staticDns: '${input.staticDns}'
[...]
```
So here's the complete cloud template that I've been working on:
```yaml
formatVersion: 1
inputs:
site:
type: string
title: Site
enum:
- BOW
- DRE
image:
type: string
title: Operating System
oneOf:
- title: Windows Server 2019
const: ws2019
default: ws2019
size:
title: Resource Size
type: string
oneOf:
- title: 'Micro [1vCPU|1GB]'
const: micro
- title: 'Tiny [1vCPU|2GB]'
const: tiny
- title: 'Small [2vCPU|2GB]'
const: small
default: small
network:
title: Network
type: string
adJoin:
title: Join to AD domain
type: boolean
default: true
staticDns:
title: Create static DNS record
type: boolean
default: false
environment:
type: string
title: Environment
oneOf:
- title: Development
const: D
- title: Testing
const: T
- title: Production
const: P
default: D
function:
type: string
title: Function Code
oneOf:
- title: Application (APP)
const: APP
- title: Desktop (DSK)
const: DSK
- title: Network (NET)
const: NET
- title: Service (SVS)
const: SVS
- title: Testing (TST)
const: TST
default: TST
app:
type: string
title: Application Code
minLength: 3
maxLength: 3
default: xxx
description:
type: string
title: Description
description: Server function/purpose
default: Testing and evaluation
poc_name:
type: string
title: Point of Contact Name
default: Jack Shephard
poc_email:
type: string
title: Point of Contact Email
default: jack.shephard@virtuallypotato.com
pattern: '^[^\s@]+@[^\s@]+\.[^\s@]+$'
ticket:
type: string
title: Ticket/Request Number
default: 4815162342
resources:
Cloud_vSphere_Machine_1:
type: Cloud.vSphere.Machine
properties:
image: '${input.image}'
flavor: '${input.size}'
site: '${input.site}'
environment: '${input.environment}'
function: '${input.function}'
app: '${input.app}'
ignoreActiveDirectory: '${!input.adJoin}'
activeDirectory:
relativeDN: '${"OU=Servers,OU=Computers,OU=" + input.site + ",OU=LAB"}'
customizationSpec: '${input.adJoin ? "vra-win-domain" : "vra-win-workgroup"}'
staticDns: '${input.staticDns}'
dnsDomain: lab.bowdre.net
poc: '${input.poc_name + " (" + input.poc_email + ")"}'
ticket: '${input.ticket}'
description: '${input.description}'
networks:
- network: '${resource.Cloud_vSphere_Network_1.id}'
assignment: static
constraints:
- tag: 'comp:${to_lower(input.site)}'
Cloud_vSphere_Network_1:
type: Cloud.vSphere.Network
properties:
networkType: existing
constraints:
- tag: 'net:${input.network}'
```
I save the template, and then also hit the "Version" button to publish a new version to the catalog:
![Releasing new version](20210803_new_template_version.png)
#### Service Broker Custom Form
I switch over to the Service Broker UI to update the custom form - but first I stop off at **Content & Policies > Content Sources**, select my Content Source, and hit the **Save & Import** button to force a sync of the cloud templates. I can then move on to the **Content & Policies > Content** section, click the 3-dot menu next to my template name, and select the option to **Customize Form**.
I'll just drag the new Schema Element called `Create static DNS record` from the Request Inputs panel and on to the form canvas. I'll drop it right below the `Join to AD domain` field:
![Adding the field to the form](20210803_updating_custom_form.png)
And then I'll hit the **Save** button so that my efforts are preserved.
That should take care of the front-end changes. Now for the back-end stuff: I need to teach vRO how to connect to my SSH host and run the PowerShell commands, [just like I tested earlier](#a-quick-test).
### The vRO solution
I will be adding the DNS action on to my existing "VM Post-Provisioning" workflow (described [here](/adding-vm-notes-and-custom-attributes-with-vra8), which gets triggered after the VM has been successfully deployed.
#### Configuration Element
But first, I'm going to go to the **Assets > Configurations** section of the Orchestrator UI and create a new Configuration Element to store variables related to the SSH host and DNS configuration.
![Create a new configuration](Go3D-gemP.png)
I'll call it `dnsConfig` and put it in my `CustomProvisioning` folder.
![Giving it a name](fJswso9KH.png)
And then I create the following variables:
| Variable | Value | Type |
| --- | --- | --- |
| `sshHost` | `win02.lab.bowdre.net` | string |
| `sshUser` | `vra` | string |
| `sshPass` | `*****` | secureString |
| `dnsServer` | `[win01.lab.bowdre.net]` | Array/string |
| `supportedDomains` | `[lab.bowdre.net]` | Array/string |
`sshHost` is my new `win02` server that I'm going to connect to via SSH, and `sshUser` and `sshPass` should explain themselves. The `dnsServer` array will tell the script which DNS servers to try to create the record on; this will just be a single server in my lab, but I'm going to construct the script to support multiple servers in case one isn't reachable. And `supported domains` will be used to restrict where I'll be creating records; again, that's just a single domain in my lab, but I'm building this solution to account for the possibility where a VM might need to be deployed on a domain where I can't create a static record in this way so I want it to fail elegantly.
Here's what the new configuration element looks like:
![Variables defined](a5gtUrQbc.png)
#### Workflow to create records
I'll need to tell my workflow about the variables held in the `dnsConfig` Configuration Element I just created. I do that by opening the "VM Post-Provisioning" workflow in the vRO UI, clicking the **Edit** button, and then switching to the **Variables** tab. I create a variable for each member of `dnsConfig`, and enable the toggle to *Bind to configuration* so that I can select the corresponding item. It's important to make sure that the variable type exactly matches what's in the configuration element so that you'll be able to pick it!
![Linking variable to config element](20210809_creating_bound_variable.png)
I repeat that for each of the remaining variables until all the members of `dnsConfig` are represented in the workflow:
![Variables added](20210809_variables_added.png)
Now we're ready for the good part: inserting a new scriptable task into the workflow schema. I'll called it `Create DNS Record` and place it directly after the `Set Notes` task. For inputs, the task will take in `inputProperties (Properties)` as well as everything from that `dnsConfig` configuration element:
![Task inputs](20210809_task_inputs.png)
And here's the JavaScript for the task:
```js
// JavaScript: Create DNS Record task
// Inputs: inputProperties (Properties), dnsServers (Array/string), sshHost (string), sshUser (string), sshPass (secureString), supportedDomains (Array/string)
// Outputs: None
var staticDns = inputProperties.customProperties.staticDns;
var hostname = inputProperties.resourceNames[0];
var dnsDomain = inputProperties.customProperties.dnsDomain;
var ipAddress = inputProperties.addresses[0];
var created = false;
// check if user requested a record to be created and if the VM's dnsDomain is in the supportedDomains array
if (staticDns == "true" && supportedDomains.indexOf(dnsDomain) >= 0) {
System.log("Attempting to create DNS record for "+hostname+"."+dnsDomain+" at "+ipAddress+"...")
// create the ssh session to the intermediary host
var sshSession = new SSHSession(sshHost, sshUser);
System.debug("Connecting to "+sshHost+"...")
sshSession.connectWithPassword(sshPass)
// loop through DNS servers in case the first one doesn't respond
for each (var dnsServer in dnsServers) {
if (created == false) {
System.debug("Using DNS Server "+dnsServer+"...")
// insert the PowerShell command to create A record
var sshCommand = 'Add-DnsServerResourceRecordA -ComputerName '+dnsServer+' -ZoneName '+dnsDomain+' -Name '+hostname+' -AllowUpdateAny -IPv4Address '+ipAddress;
System.debug("sshCommand: "+sshCommand)
// run the command and check the result
sshSession.executeCommand(sshCommand, true)
var result = sshSession.exitCode;
if (result == 0) {
System.log("Successfully created DNS record!")
// make a note that it was successful so we don't repeat this unnecessarily
created = true;
}
}
}
sshSession.disconnect()
if (created == false) {
System.warn("Error! Unable to create DNS record.")
}
} else {
System.log("Not trying to do DNS")
}
```
Now I can just save the workflow, and I'm done! - with this part. Of course, being able to *create* a static record is just one half of the fight; I also need to make sure that vRA will be able to clean up these static records when a deployment gets deleted.
#### Workflow to delete records
I haven't previously created any workflows that fire on deployment removal, so I'll create a new one and call it `VM Deprovisioning`:
![New workflow](20210811_new_workflow.png)
This workflow only needs a single input (`inputProperties (Properties)`) so it can receive information about the deployment from vRA:
![Workflow input](20210811_inputproperties.png)
I'll also need to bind in the variables from the `dnsConfig` element as before:
![Workflow variables](20210812_deprovision_variables.png)
The schema will include a single scriptable task:
![Delete DNS Record task](20210812_delete_dns_record_task.png)
And it's going to be *pretty damn similar* to the other one:
```js
// JavaScript: Delete DNS Record task
// Inputs: inputProperties (Properties), dnsServers (Array/string), sshHost (string), sshUser (string), sshPass (secureString), supportedDomains (Array/string)
// Outputs: None
var staticDns = inputProperties.customProperties.staticDns;
var hostname = inputProperties.resourceNames[0];
var dnsDomain = inputProperties.customProperties.dnsDomain;
var ipAddress = inputProperties.addresses[0];
var deleted = false;
// check if user requested a record to be created and if the VM's dnsDomain is in the supportedDomains array
if (staticDns == "true" && supportedDomains.indexOf(dnsDomain) >= 0) {
System.log("Attempting to remove DNS record for "+hostname+"."+dnsDomain+" at "+ipAddress+"...")
// create the ssh session to the intermediary host
var sshSession = new SSHSession(sshHost, sshUser);
System.debug("Connecting to "+sshHost+"...")
sshSession.connectWithPassword(sshPass)
// loop through DNS servers in case the first one doesn't respond
for each (var dnsServer in dnsServers) {
if (deleted == false) {
System.debug("Using DNS Server "+dnsServer+"...")
// insert the PowerShell command to delete A record
var sshCommand = 'Remove-DnsServerResourceRecord -ComputerName '+dnsServer+' -ZoneName '+dnsDomain+' -RRType A -Name '+hostname+' -Force';
System.debug("sshCommand: "+sshCommand)
// run the command and check the result
sshSession.executeCommand(sshCommand, true)
var result = sshSession.exitCode;
if (result == 0) {
System.log("Successfully deleted DNS record!")
// make a note that it was successful so we don't repeat this unnecessarily
deleted = true;
}
}
}
sshSession.disconnect()
if (deleted == false) {
System.warn("Error! Unable to delete DNS record.")
}
} else {
System.log("No need to clean up DNS.")
}
```
Since this is a new workflow, I'll also need to head back to **Cloud Assembly > Extensibility > Subscriptions** and add a new subscription to call it when a deployment gets deleted. I'll call it "VM Deprovisioning", assign it to the "Compute Post Removal" Event Topic, and link it to my new "VM Deprovisioning" workflow. I *could* use the Condition option to filter this only for deployments which had a static DNS record created, but I'll later want to use this same workflow for other cleanup tasks so I'll just save it as is for now.
![VM Deprovisioning subscription](20210812_deprovisioning_subscription.png)
### Testing
Now I can (finally) fire off a quick deployment to see if all this mess actually works:
![Test deploy request](20210812_test_deploy_request.png)
Once the deployment completes, I go back into vRO, find the most recent item in the **Workflow Runs** view, and click over to the **Logs** tab to see how I did:
![Workflow success!](20210813_workflow_success.png)
And I can run a quick query to make sure that name actually resolves:
```shell
dig +short bow-ttst-xxx023.lab.bowdre.net A
172.16.30.10
```
It works!
Now to test the cleanup. For that, I'll head back to Service Broker, navigate to the **Deployments** tab, find my deployment, click the little three-dot menu button, and select the **Delete** option:
![Deleting the deployment](20210813_delete_deployment.png)
Again, I'll check the **Workflow Runs** in vRO to see that the deprovisioning task completed successfully:
![VM Deprovisioning workflow](20210813_workflow_deletion.png)
And I can `dig` a little more to make sure the name doesn't resolve anymore:
```shell
dig +short bow-ttst-xxx023.lab.bowdre.net A
```
It *really* works!
### Conclusion
So there you have it - how I've got vRA/vRO able to create and delete static DNS records as needed, using a Windows SSH host as an intermediary. Cool, right?

Binary file not shown.

After

Width:  |  Height:  |  Size: 136 KiB

View file

@ -0,0 +1,83 @@
---
date: "2020-09-22T08:34:30Z"
thumbnail: 8p-PSHx1R.png
usePageBundles: true
tags:
- docker
- windows
- wsl
- containers
title: Docker on Windows 10 with WSL2
---
Microsoft's Windows Subsystem for Linux (WSL) 2 [was recently updated](https://devblogs.microsoft.com/commandline/wsl-2-support-is-coming-to-windows-10-versions-1903-and-1909/) to bring support for less-bleeding-edge Windows 10 versions (like 1903 and 1909). WSL2 is a big improvement over the first iteration (particularly with [better Docker support](https://www.docker.com/blog/docker-desktop-wsl-2-backport-update/)) so I was really looking forward to getting WSL2 loaded up on my work laptop.
Here's how.
### WSL2
#### Step Zero: Prereqs
You'll need Windows 10 1903 build 18362 or newer (on x64). You can check by running `ver` from a Command Prompt:
```powershell
C:\> ver
Microsoft Windows [Version 10.0.18363.1082]
```
We're interested in that third set of numbers. 18363 is bigger than 18362 so we're good to go!
#### Step One: Enable the WSL feature
*(Not needed if you've already been using WSL1.)*
You can do this by dropping the following into an elevated Powershell prompt:
```powershell
dism.exe /online /enable-feature /featurename:Microsoft-Windows-Subsystem-Linux /all /norestart
```
#### Step Two: Enable the Virtual Machine Platform feature
Drop this in an elevated Powershell:
```powershell
dism.exe /online /enable-feature /featurename:VirtualMachinePlatform /all /norestart
```
And then reboot (this is still Windows, after all).
#### Step Three: Install the WSL2 kernel update package
Download it from [here](https://wslstorestorage.blob.core.windows.net/wslblob/wsl_update_x64.msi), and double-click the downloaded file to install it.
#### Step Four: Set WSL2 as your default
Open a Powershell window and run:
```powershell
wsl --set-default-version 2
```
#### Step Five: Install a Linux distro, or upgrade an existing one
If you're brand new to this WSL thing, head over to the [Microsoft Store](https://aka.ms/wslstore) and download your favorite Linux distribution. Once it's installed, launch it and you'll be prompted to set up a Linux username and password.
If you've already got a WSL1 distro installed, first run `wsl -l -v` in Powershell to make sure you know the distro name:
```powershell
PS C:\Users\jbowdre> wsl -l -v
NAME STATE VERSION
* Debian Running 2
```
And then upgrade the distro to WSL2 with `wsl --set-version <distro_name> 2`:
```powershell
PS C:\Users\jbowdre> wsl --set-version Debian 2
Conversion in progress, this may take a few minutes...
```
Cool!
### Docker
#### Step One: Download
Download Docker Desktop for Windows from [here](https://hub.docker.com/editions/community/docker-ce-desktop-windows/), making sure to grab the "Edge" version since it includes support for the backported WSL2 bits.
#### Step Two: Install
Run the installer, and make sure to tick the box for installing the WSL2 engine.
#### Step Three: Configure Docker Desktop
Launch Docker Desktop from the Start menu, and you should be presented with this friendly prompt:
![Great news! We're supported.](lY2FTflbK.png)
Hit that big friendly "gimme WSL2" button. Then open the Docker Settings from the system tray, and make sure that **General > Use the WSL 2 based engine** is enabled. Now navigate to **Resources > WSL Integration**, confirm that **Enable integration with my default WSL distro** is enabled as well. Smash the "Apply & Restart" button if you've made any changes.
### Test it!
Fire up a WSL session and confirm that everything is working with `docker run hello-world`:
![Hello, world!](8p-PSHx1R.png)
It's beautiful!

Binary file not shown.

After

Width:  |  Height:  |  Size: 55 KiB

View file

@ -0,0 +1,67 @@
---
title: "Enable Tanzu CLI Auto-Completion in bash and zsh" # Title of the blog post.
date: 2022-02-01T08:34:47-06:00 # Date of post creation.
# lastmod: 2022-02-01T08:34:47-06:00 # Date when last modified
description: "How to configure your Linux shell to help you do the Tanzu" # Description used for search engine.
featured: false # Sets if post is a featured post, making appear on the home page side bar.
draft: false # Sets whether to render this page. Draft of true will not be rendered.
toc: false # Controls if a table of contents should be generated for first-level links automatically.
usePageBundles: true
# menu: main
# featureImage: "tanzu-completion.png" # Sets featured image on blog post.
# featureImageAlt: 'Description of image' # Alternative text for featured image.
# featureImageCap: 'This is the featured image.' # Caption (optional).
thumbnail: "tanzu-completion.png" # Sets thumbnail image appearing inside card on homepage.
# shareImage: "share.png" # Designate a separate image for social media sharing.
codeLineNumbers: false # Override global value for showing of line numbers within code block.
series: Tips
tags:
- vmware
- linux
- tanzu
- kubernetes
- shell
comment: true # Disable comment if false.
---
Lately I've been spending some time [getting more familiar](/tanzu-community-edition-k8s-homelab/) with VMware's [Tanzu Community Edition](https://tanzucommunityedition.io/) Kubernetes distribution, but I'm still not quite familiar enough with the `tanzu` command line. If only there were a better way for me to discover the available commands for a given context and help me type them correctly...
Oh, but there is! You see, one of the available Tanzu commands is `tanzu completion [shell]`, which will spit out the necessary code to generate handy context-based auto-completions appropriate for the shell of your choosing (provided that you choose either `bash` or `zsh`, that is).
Running `tanzu completion --help` will tell you what's needed, and you can just copy/paste the commands appropriate for your shell:
```shell
# Bash instructions:
## Load only for current session:
source <(tanzu completion bash)
## Load for all new sessions:
tanzu completion bash > $HOME/.tanzu/completion.bash.inc
printf "\n# Tanzu shell completion\nsource '$HOME/.tanzu/completion.bash.inc'\n" >> $HOME/.bash_profile
# Zsh instructions:
## Load only for current session:
source <(tanzu completion zsh)
## Load for all new sessions:
echo "autoload -U compinit; compinit" >> ~/.zshrc
tanzu completion zsh > "${fpath[1]}/_tanzu"
```
So to get the completions to load automatically whenever you start a `bash` shell, run:
```shell
tanzu completion bash > $HOME/.tanzu/completion.bash.inc
printf "\n# Tanzu shell completion\nsource '$HOME/.tanzu/completion.bash.inc'\n" >> $HOME/.bash_profile
```
For a `zsh` shell, it's:
```shell
echo "autoload -U compinit; compinit" >> ~/.zshrc
tanzu completion zsh > "${fpath[1]}/_tanzu"
```
And that's it! The next time you open a shell (or `source` your relevant profile), you'll be able to `[TAB]` your way through the Tanzu CLI!
![Tanzu CLI completion in zsh](tanzu-completion.gif)

Binary file not shown.

After

Width:  |  Height:  |  Size: 334 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 120 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 47 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 112 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 122 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 41 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 75 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 80 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 61 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 103 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 124 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 139 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 116 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 66 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 62 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 125 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 350 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 51 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 76 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 107 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 68 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 35 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 121 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 74 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 70 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 228 KiB

View file

@ -0,0 +1,416 @@
---
title: "ESXi ARM Edition on the Quartz64 SBC" # Title of the blog post.
date: 2022-04-23 # Date of post creation.
lastmod: 2022-12-14
description: "Getting started with the experimental ESXi Arm Edition fling to run a VMware hypervisor on the PINE64 Quartz64 single-board computer, and installing a Tailscale node on Photon OS to facilitate improved remote access to my home network." # Description used for search engine.
featured: true # Sets if post is a featured post, making appear on the home page side bar.
draft: false # Sets whether to render this page. Draft of true will not be rendered.
toc: true # Controls if a table of contents should be generated for first-level links automatically.
usePageBundles: true
# menu: main
featureImage: "quartz64.jpg" # Sets featured image on blog post.
# featureImageAlt: 'Description of image' # Alternative text for featured image.
# featureImageCap: 'This is the featured image.' # Caption (optional).
thumbnail: "quartz64.jpg" # Sets thumbnail image appearing inside card on homepage.
# shareImage: "share.png" # Designate a separate image for social media sharing.
codeLineNumbers: false # Override global value for showing of line numbers within code block.
series: Projects
tags:
- vmware
- linux
- chromeos
- homelab
- tailscale
- photon
- vpn
comment: true # Disable comment if false.
---
{{% notice info "ESXi-ARM Fling v1.10 Update" %}}
On July 20, 2022, VMware released a [major update](https://blogs.vmware.com/arm/2022/07/20/1-10/) for the ESXi-ARM Fling. Among [other fixes and improvements](https://flings.vmware.com/esxi-arm-edition#changelog), this version enables **in-place ESXi upgrades** and [adds support for the Quartz64's **on-board NIC**](https://twitter.com/jmcwhatever/status/1549935971822706688). To update, I:
1. Wrote the new ISO installer to another USB drive.
2. Attached the installer drive to the USB hub, next to the existing ESXi drive.
3. Booted the installer and selected to upgrade ESXi on the existing device.
4. Powered-off post-install, unplugged the hub, and attached the ESXi drive directly to the USB2 port on the Quart64.
5. Connected the ethernet cable to the onboard NIC.
6. Booted to ESXi.
7. Once booted, I used the DCUI to (re)configure the management network and activate the onboard network adapter.
Now I've got directly-attached USB storage, and the onboard NIC provides gigabit connectivity. I've made a few tweaks to the rest of the article to reflect the lifting of those previous limitations.
{{% /notice %}}
Up until this point, [my homelab](/vmware-home-lab-on-intel-nuc-9/) has consisted of just a single Intel NUC9 ESXi host running a bunch of VMs. It's served me well but lately I've been thinking that it would be good to have an additional host for some of my workloads. In particular, I'd like to have a [Tailscale node](/secure-networking-made-simple-with-tailscale/) on my home network which _isn't_ hosted on the NUC so that I can patch ESXi remotely without cutting off my access. I appreciate the small footprint of the NUC so I'm not really interested in a large "grown-up" server at this time. So for now I thought it might be fun to experiment with [VMware's ESXi on ARM fling](https://flings.vmware.com/esxi-arm-edition) which makes it possible to run a full-fledged VMWare hypervisor on a Raspbery Pi.
Of course, I decided to embark upon this project at a time when Raspberry Pis are basically impossible to get. So instead I picked up a [PINE64 Quartz64](https://wiki.pine64.org/wiki/Quartz64) single-board computer (SBC) which seems like a potentially very-capable piece of hardware.... but there is a prominent warning at the bottom of the [store page](https://pine64.com/product/quartz64-model-a-8gb-single-board-computer/):
{{% notice warning "Be Advised" %}}
"The Quartz64 SBC still in early development stage, only suitable for developers and advanced users wishing to contribute to early software development. Both mainline and Rockchips BSP fork of Linux have already been booted on the platform and development is proceeding quickly, but it will be months before end-users and industry partners can reliably deploy it. If you need a single board computer for a private or industrial application today, we encourage you to choose a different board from our existing lineup or wait a few months until Quartz64 software reaches a sufficient degree of maturity."
{{% /notice %}}
More specifically, for my use case there will be a number of limitations (at least for now - this SBC is still pretty new to the market so hopefully support will be improving further over time):
- ~~The onboard NIC is not supported by ESXi.~~[^v1.10]
- Onboard storage (via eMMC, eSATA, or PCIe) is not supported.
- The onboard microSD slot is only used for loading firmware on boot, not for any other storage.
- Only two (of the four) USB ports are documented to work reliably.
- Of the remaining two ports, the lower USB3 port [shouldn't be depended upon either](https://wiki.pine64.org/wiki/Quartz64_Development#Confirmed_Broken) so I'm really just stuck with a single USB2 interface ~~which will need to handle both networking and storage~~[^v1.10].[^usb3]
All that is to say that (as usual) I'll be embarking upon this project in Hard Mode - and I'll make it extra challenging (as usual) by doing all of the work from a Chromebook. In any case, here's how I managed to get ESXi running on the the Quartz64 SBC and then deploy a small workload.
[^usb3]: Jared McNeill, the maintainer of the firmware image I'm using *just* [pushed a commit](https://github.com/jaredmcneill/quartz64_uefi/commit/4bda76e9fce5ed153ac49fa9d51ff34e5dd56d52) which sounds like it may address this flaky USB3 issue but that was after I had gotten everything else working as described below. I'll check that out once a new release gets published.
[^v1.10]: Fixed in the v1.10 release.
### Bill of Materials
Let's start with the gear (hardware and software) I needed to make this work:
| Hardware | Purpose |
| --- | --- |
| [PINE64 Quartz64 Model-A 8GB Single Board Computer](https://pine64.com/product/quartz64-model-a-8gb-single-board-computer/) | kind of the whole point |
| [ROCKPro64 12V 5A US Power Supply](https://pine64.com/product/rockpro64-12v-5a-us-power-supply/) | provies power for the the SBC |
| [Serial Console “Woodpecker” Edition](https://pine64.com/product/serial-console-woodpecker-edition/) | allows for serial console access |
| [Google USB-C Adapter](https://www.amazon.com/dp/B071G6NLHJ/) | connects the console adapter to my Chromebook |
| [Sandisk 64GB Micro SD Memory Card](https://www.amazon.com/dp/B00M55C1I2) | only holds the firmware; a much smaller size would be fine |
| [Monoprice USB-C MicroSD Reader](https://www.amazon.com/dp/B00YQM8352/) | to write firmware to the SD card from my Chromebook |
| [Samsung MUF-256AB/AM FIT Plus 256GB USB 3.1 Drive](https://www.amazon.com/dp/B07D7Q41PM) | ESXi boot device and local VMFS datastore |
| ~~[Cable Matters 3 Port USB 3.0 Hub with Ethernet](https://www.amazon.com/gp/product/B01J6583NK)~~ | ~~for network connectivity and to host the above USB drive~~[^v1.10] |
| [3D-printed open enclosure for QUARTZ64](https://www.thingiverse.com/thing:5308499) | protect the board a little bit while allowing for plenty of passive airflow |
| Downloads | Purpose |
| --- | --- |
| [ESXi ARM Edition](https://customerconnect.vmware.com/downloads/get-download?downloadGroup=ESXI-ARM) (v1.10) | hypervisor |
| [Tianocore EDK II firmware for Quartz64](https://github.com/jaredmcneill/quartz64_uefi/releases) (2022-07-20) | firmare image |
| [Chromebook Recovery Utility](https://chrome.google.com/webstore/detail/chromebook-recovery-utili/pocpnlppkickgojjlmhdmidojbmbodfm) | easy way to write filesystem images to external media |
| [Beagle Term](https://chrome.google.com/webstore/detail/beagle-term/gkdofhllgfohlddimiiildbgoggdpoea) | for accessing the Quartz64 serial console |
### Preparation
#### Firmware media
The very first task is to write the required firmware image (download [here](https://github.com/jaredmcneill/quartz64_uefi/releases)) to a micro SD card. I used a 64GB card that I had lying around but you could easily get by with a *much* smaller one; the firmware image is tiny, and the card can't be used for storing anything else. Since I'm doing this on a Chromebook, I'll be using the [Chromebook Recovery Utility (CRU)](https://chrome.google.com/webstore/detail/chromebook-recovery-utili/pocpnlppkickgojjlmhdmidojbmbodfm) for writing the images to external storage as described [in another post](/burn-an-iso-to-usb-with-the-chromebook-recovery-utility/).
After downloading [`QUARTZ64_EFI.img.gz`](https://github.com/jaredmcneill/quartz64_uefi/releases/download/2022-07-20/QUARTZ64_EFI.img.gz), I need to get it into a format recognized by CRU and, in this case, that means extracting the gzipped archive and then compressing the `.img` file into a standard `.zip`:
```
gunzip QUARTZ64_EFI.img.gz
zip QUARTZ64_EFI.img.zip QUARTZ64_EFI.img
```
I can then write it to the micro SD card by opening CRU, clicking on the gear icon, and selecting the *Use local image* option.
![Writing the firmware image](writing_firmware.png)
#### ESXi installation media
I'll also need to prepare the ESXi installation media (download [here](https://customerconnect.vmware.com/downloads/get-download?downloadGroup=ESXI-ARM)). For that, I'll be using a 256GB USB drive. Due to the limited storage options on the Quartz64, I'll be installing ESXi onto the same drive I use to boot the installer so, in this case, the more storage the better. By default, ESXi 7.0 will consume up to 128GB for the new `ESX-OSData` partition; whatever is leftover will be made available as a VMFS datastore. That could be problematic given the unavailable/flaky USB support of the Quartz64. (While you *can* install ESXi onto a smaller drive, down to about ~20GB, the lack of additional storage on this hardware makes it pretty important to take advantage of as much space as you can.)
In any case, to make the downloaded `VMware-VMvisor-Installer-7.0-20133114.aarch64.iso` writeable with CRU all I need to do is add `.bin` to the end of the filename:
```
mv VMware-VMvisor-Installer-7.0-20133114.aarch64.iso{,.bin}
```
Then it's time to write the image onto the USB drive:
![Writing the ESXi installer image](writing_esxi.png)
#### Console connection
I'll need to use the Quartz64 serial console interface and ["Woodpecker" edition console USB adapter](https://pine64.com/product/serial-console-woodpecker-edition/) to interact with the board until I get ESXi installed and can connect to it with the web interface or SSH. The adapter comes with a short breakout cable, and I connect it thusly:
| Quartz64 GPIO pin | Console adapter pin | Wire color |
| --- | --- | --- |
| 6 | `GND` | Brown |
| 8 | `RXD` | Red |
| 10 | `TXD` | Orange |
I leave the yellow wire dangling free on both ends since I don't need a `+V` connection for the console to work.
![Console connection](console_connection.jpg)
To verify that I've got things working, I go ahead and pop the micro SD card containing the firmware into its slot on the bottom side of the Quartz64 board, connect the USB console adapter to my Chromebook, and open the [Beagle Term](https://chrome.google.com/webstore/detail/beagle-term/gkdofhllgfohlddimiiildbgoggdpoea) app to set up the serial connection.
I'll need to use these settings for the connection (which are the defaults selected by Beagle Term):
| Setting | Value |
| -- | --- |
| Port | `/dev/ttyUSB0` |
| Bitrate | `115200` |
| Data Bit | `8 bit` |
| Parity | `none` |
| Stop Bit | `1` |
| Flow Control | `none` |
![Beagle Term settings](beagle_term_settings.png)
I hit **Connect** and then connect the Quartz64's power supply. I watch as it loads the firmware and then launches the BIOS menu:
![BIOS menu](bios.png)
### Host creation
#### ESXi install
Now that I've got everything in order I can start the install. A lot of experimentation on my part confirmed the sad news about the USB ports: of the four USB ports, only the top-right USB2 port works reliably for me. So I connect my ~~USB NIC+hub to that port, and plug in my 256GB drive to the hub~~[^v1.10] 256GB USB drive there. This isn't ideal from a performance aspect, of course, but slow storage is more useful than no storage.
On that note, remember what I mentioned earlier about how the ESXi installer would want to fill up ~128GB worth of whatever drive it targets? The ESXi ARM instructions say that you can get around that by passing the `autoPartitionOSDataSize` advanced option to the installer by pressing `[Shift] + O` in the ESXi bootloader, but the Quartz64-specific instructions say that you can't do that with this board since only the serial console is available... It turns out this is a (happy) lie.
I hooked up a monitor to the board's HDMI port and a USB keyboard to a free port on the hub and verified that the keyboard let me maneuver through the BIOS menu. From here, I hit the **Reset** button on the Quartz64 to restart it and let it boot from the connected USB drive. When I got to the ESXi pre-boot countdown screen, I pressed `[Shift] + O` as instructed and added `autoPartitionOSDataSize=8192` to the boot options. This limits the size of the new-for-ESXi7 ESX-OSData VMFS-L volume to 8GB and will give me much more space for the local datastore.
Beyond that it's a fairly typical ESXi install process:
![Hi, welcome to the ESXi for ARM installer. I'll be your UI this evening.](esxi_install_1.png)
![Just to be sure, I'm going to clobber everything on this USB drive.](esxi_install_2.png)
![Hold on to your butts, here we go!](esxi_install_3.png)
![Whew, we made it!](esxi_install_4.png)
#### Initial configuration
After the installation completed, I rebooted the host and watched for the Direct Console User Interface (DCUI) to come up:
![ESXi DCUI](dcui.png)
I hit `[F2]` and logged in with the root credentials to get to the System Customization menu:
![DCUI System Customization](dcui_system_customization.png)
The host automatically received an IP issued by DHCP but I'd like for it to instead use a static IP. I'll also go ahead and configure the appropriate DNS settings.
![Setting the IP address](dcui_ip_address.png)
![Configuring DNS settings](dcui_dns.png)
I also create the appropriate matching `A` and `PTR` records in my local DNS, and (after bouncing the management network) I can access the ESXi Embedded Host Client at `https://quartzhost.lab.bowdre.net`:
![ESXi Embedded Host Client login screen](embedded_host_client_login.png)
![Summary view of my new host!](embedded_host_client_summary.png)
That's looking pretty good... but what's up with that date and time? Time has kind of lost all meaning in the last couple of years but I'm *reasonably* certain that January 1, 2001 was at least a few years ago. And I know from past experience that incorrect host time will prevent it from being successfully imported to a vCenter inventory.
Let's clear that up by enabling the Network Time Protocol (NTP) service on this host. I'll do that by going to **Manage > System > Time & Date** and clicking the **Edit NTP Settings** button. I don't run a local NTP server so I'll point it at `pool.ntp.org` and set the service to start and stop with the host:
![NTP configuration](ntp_configuration.png)
Now I hop over to the **Services** tab, select the `ntpd` service, and then click the **Start** button there. Once it's running, I then *restart* `ntpd` to help encourage the system to update the time immediately.
![Starting the NTP service](services.png)
Once the service is started I can go back to **Manage > System > Time & Date**, click the **Refresh** button, and confirm that the host has been updated with the correct time:
![Correct time!](correct_time.png)
With the time sorted, I'm just about ready to join this host to my vCenter, but first I'd like to take a look at the storage situation - after all, I did jump through those hoops with the installer to make sure that I would wind up with a useful local datastore. Upon going to **Storage > More storage > Devices** and clicking on the single listed storage device, I can see in the Partition Diagram that the ESX-OSData VMFS-L volume was indeed limited to 8GB, and the free space beyond that was automatically formatted as a VMFS datastore:
![Reviewing the partition diagram](storage_device.png)
And I can also take a peek at that local datastore:
![Local datastore](storage_datastore.png)
With 200+ gigabytes of free space on the datastore I should have ample room for a few lightweight VMs.
#### Adding to vCenter
Alright, let's go ahead and bring the new host into my vCenter environment. That starts off just like any other host, by right-clicking an inventory location in the *Hosts & Clusters* view and selecting **Add Host**.
![Starting the process](add_host.png)
![Reviewing the host details](add_host_confirm.png)
![Successfully added to the vCenter](host_added.png)
Success! I've now got a single-board hypervisor connected to my vCenter. Now let's give that host a workload.[^workloads]
[^workloads]: Hosts *love* workloads.
### Workload creation
As I mentioned earlier, my initial goal is to deploy a Tailscale node on my new host so that I can access my home network from outside of the single-host virtual lab environment. I've become a fan of using VMware's [Photon OS](https://vmware.github.io/photon/) so I'll get a VM deployed and then install the Tailscale agent.
#### Deploying Photon OS
VMware provides Photon in a few different formats, as described on the [download page](https://github.com/vmware/photon/wiki/Downloading-Photon-OS). I'm going to use the "OVA with virtual hardware v13 arm64" version so I'll kick off that download of `photon_uefi.ova`. I'm actually going to download that file straight to my `deb01` Linux VM:
```shell
wget https://packages.vmware.com/photon/4.0/Rev2/ova/photon_uefi.ova
```
and then spawn a quick Python web server to share it out:
```shell
python3 -m http.server
Serving HTTP on 0.0.0.0 port 8000 (http://0.0.0.0:8000/) ...
```
That will let me deploy from a resource already inside my lab network instead of transferring the OVA from my laptop. So now I can go back to my vSphere Client and go through the steps to **Deploy OVF Template** to the new host, and I'll plug in the URL `http://deb01.lab.bowdre.net:8000/photon_uefi.ova`:
![Deploying a template from URL](deploy_from_url.png)
I'll name it `pho01` and drop it in an appropriate VM folder:
![Naming the new VM](name_vm.png)
And place it on the new Quartz64 host:
![Host placement](vm_placement.png)
The rest of the OVF deployment is basically just selecting the default options and clicking through to finish it. And then once it's deployed, I'll go ahead and power on the new VM.
![The newly-created Photon VM](new_vm.png)
#### Configuring Photon
There are just a few things I'll want to configure on this VM before I move on to installing Tailscale, and I'll start out simply by logging in with the remote console.
{{% notice info "Default credentials" %}}
The default password for Photon's `root` user is `changeme`. You'll be forced to change that at first login.
{{% /notice %}}
![First login, and the requisite password change](first_login.png)
Now that I'm in, I'll set the hostname appropriately:
```bash
hostnamectl set-hostname pho01
```
For now, the VM pulled an IP from DHCP but I would like to configure that statically instead. To do that, I'll create a new interface file:
```bash
cat > /etc/systemd/network/10-static-en.network << "EOF"
[Match]
Name = eth0
[Network]
Address = 192.168.1.17/24
Gateway = 192.168.1.1
DNS = 192.168.1.5
DHCP = no
IPForward = yes
EOF
chmod 644 /etc/systemd/network/10-static-en.network
systemctl restart systemd-networkd
```
I'm including `IPForward = yes` to [enable IP forwarding](https://tailscale.com/kb/1104/enable-ip-forwarding/) for Tailscale.
With networking sorted, it's probably a good idea to check for and apply any available updates:
```bash
tdnf update -y
```
I'll also go ahead and create a normal user account (with sudo privileges) for me to use:
```bash
useradd -G wheel -m john
passwd john
```
Now I can use SSH to connect to the VM and ditch the web console:
```bash
ssh pho01.lab.bowdre.net
Password:
john@pho01 [ ~ ]$ sudo whoami
We trust you have received the usual lecture from the local System
Administrator. It usually boils down to these three things:
#1) Respect the privacy of others.
#2) Think before you type.
#3) With great power comes great responsibility.
[sudo] password for john
root
```
Looking good! I'll now move on to the justification[^justification] for this entire exercise:
[^justification]: Entirely arbitrary and fabricated justification.
#### Installing Tailscale
If I *weren't* doing this on hard mode, I could use Tailscale's [install script](https://tailscale.com/download) like I do on every other Linux system. Hard mode is what I do though, and the installer doesn't directly support Photon OS. I'll instead consult the [manual install instructions](https://tailscale.com/download/linux/static) which tell me to download the appropriate binaries from [https://pkgs.tailscale.com/stable/#static](https://pkgs.tailscale.com/stable/#static). So I'll grab the link for the latest `arm64` build and pull the down to the VM:
```bash
curl https://pkgs.tailscale.com/stable/tailscale_1.22.2_arm64.tgz --output tailscale_arm64.tgz
```
Then I can unpack it:
```bash
sudo tdnf install tar
tar xvf tailscale_arm64.tgz
cd tailscale_1.22.2_arm64/
```
So I've got the `tailscale` and `tailscaled` binaries as well as some sample service configs in the `systemd` directory:
```bash
john@pho01 [ ~/tailscale_1.22.2_arm64 ]$
.:
total 32288
drwxr-x--- 2 john users 4096 Mar 18 02:44 systemd
-rwxr-x--- 1 john users 12187139 Mar 18 02:44 tailscale
-rwxr-x--- 1 john users 20866538 Mar 18 02:44 tailscaled
./systemd:
total 8
-rw-r----- 1 john users 287 Mar 18 02:44 tailscaled.defaults
-rw-r----- 1 john users 674 Mar 18 02:44 tailscaled.service
```
Dealing with the binaries is straight-forward. I'll drop them into `/usr/bin/` and `/usr/sbin/` (respectively) and set the file permissions:
```bash
sudo install -m 755 tailscale /usr/bin/
sudo install -m 755 tailscaled /usr/sbin/
```
Then I'll descend to the `systemd` folder and see what's up:
```bash
john@pho01 [ ~/tailscale_1.22.2_arm64/ ]$ cd systemd/
john@pho01 [ ~/tailscale_1.22.2_arm64/systemd ]$ cat tailscaled.defaults
# Set the port to listen on for incoming VPN packets.
# Remote nodes will automatically be informed about the new port number,
# but you might want to configure this in order to set external firewall
# settings.
PORT="41641"
# Extra flags you might want to pass to tailscaled.
FLAGS=""
john@pho01 [ ~/tailscale_1.22.2_arm64/systemd ]$ cat tailscaled.service
[Unit]
Description=Tailscale node agent
Documentation=https://tailscale.com/kb/
Wants=network-pre.target
After=network-pre.target NetworkManager.service systemd-resolved.service
[Service]
EnvironmentFile=/etc/default/tailscaled
ExecStartPre=/usr/sbin/tailscaled --cleanup
ExecStart=/usr/sbin/tailscaled --state=/var/lib/tailscale/tailscaled.state --socket=/run/tailscale/tailscaled.sock --port $PORT $FLAGS
ExecStopPost=/usr/sbin/tailscaled --cleanup
Restart=on-failure
RuntimeDirectory=tailscale
RuntimeDirectoryMode=0755
StateDirectory=tailscale
StateDirectoryMode=0700
CacheDirectory=tailscale
CacheDirectoryMode=0750
Type=notify
[Install]
WantedBy=multi-user.target
```
`tailscaled.defaults` contains the default configuration that will be referenced by the service, and `tailscaled.service` tells me that it expects to find it at `/etc/defaults/tailscaled`. So I'll copy it there and set the perms:
```bash
sudo install -m 644 tailscaled.defaults /etc/defaults/tailscaled
```
`tailscaled.service` will get dropped in `/usr/lib/systemd/system/`:
```bash
sudo install -m 644 tailscaled.service /usr/lib/systemd/system/
```
Then I'll enable the service and start it:
```bash
sudo systemctl enable tailscaled.service
sudo systemctl start tailscaled.service
```
And finally log in to Tailscale, including my `tag:home` tag for [ACL purposes](/secure-networking-made-simple-with-tailscale/#acls) and a route advertisement for my home network so that my other Tailscale nodes can use this one to access other devices as well:
```bash
sudo tailscale up --advertise-tags "tag:home" --advertise-route "192.168.1.0/24"
```
That will return a URL I can use to authenticate, and I'll then able to to view and manage the new Tailscale node from the `login.tailscale.com` admin portal:
![Success!](new_tailscale_node.png)
You might remember [from last time](/secure-networking-made-simple-with-tailscale/#subnets-and-exit-nodes) that the "Subnets (!)" label indicates that this node is attempting to advertise a subnet route but that route hasn't yet been accepted through the admin portal. You may also remember that the `192.168.1.0/24` subnet is already being advertised by my `vyos` node:[^hassos]
![Actively-routed subnets show up black, while advertised-but-not-currently-routed subnets appear grey](advertised_subnets.png)
Things could potentially get messy if I have two nodes advertising routes for the same subnet[^failover] so I'm going to use the admin portal to disable that route on `vyos` before enabling it for `pho01`. I'll let `vyos` continue to route the `172.16.0.0/16` subnet (which only exists inside the NUC's vSphere environment after all) and it can continue to function as an Exit Node as well.
![Disabling the subnet on vyos](disabling_subnet_on_vyos.png)
![Enabling the subnet on pho01](enabling_subnet_on_pho01.png)
![Updated subnets](updated_subnets.png)
Now I can remotely access the VM (and thus my homelab!) from any of my other Tailscale-enrolled devices!
[^hassos]: The [Tailscale add-on for Home Assistant](https://github.com/hassio-addons/addon-tailscale) also tries to advertise its subnets by default, but I leave that disabled in the admin portal as well.
[^failover]: Tailscale does offer a [subnet router failover feature](https://tailscale.com/kb/1115/subnet-failover/) but it is only available starting on the [Business ($15/month) plan](https://tailscale.com/pricing/) and not the $48/year Personal Pro plan that I'm using.
### Conclusion
I actually received the Quartz64 waay back on March 2nd, and it's taken me until this week to get all the pieces in place and working the way I wanted.
{{< tweet user="johndotbowdre" id="1499194756148125701" >}}
As is so often the case, a lot of time and effort would have been saved if I had RTFM'd[^rtfm] before diving in to the deep end. I definitely hadn't anticipated all the limitations that would come with the Quartz64 SBC before ordering mine. Now that it's done, though, I'm pretty pleased with the setup, and I feel like I learned quite a bit along the way. I keep reminding myself that this is still a very new hardware platform. I'm excited to see how things improve with future development efforts.
[^rtfm]: Read The *Friendly* Manual. Yeah.

Binary file not shown.

After

Width:  |  Height:  |  Size: 80 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 65 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 159 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 120 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 106 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 230 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 87 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 73 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 121 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 127 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 32 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 91 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 85 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 160 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 153 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 162 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 79 KiB

View file

@ -0,0 +1,120 @@
---
series: vRA8
date: "2021-11-05T00:00:00Z"
thumbnail: 20211105_ssc_403.png
usePageBundles: true
tags:
- vmware
- vra
- lcm
- salt
- openssl
- certs
title: Fixing 403 error on SaltStack Config 8.6 integrated with vRA and vIDM
---
I've been wanting to learn a bit more about [SaltStack Config](https://www.vmware.com/products/vrealize-automation/saltstack-config.html) so I recently deployed SSC 8.6 to my environment (using vRealize Suite Lifecycle Manager to do so as [described here](https://cosmin.gq/2021/02/02/deploying-saltstack-config-via-lifecycle-manager-in-a-vra-environment/)). I selected the option to integrate with my pre-existing vRA and vIDM instances so that I wouldn't have to manage authentication directly since I recall that the LDAP authentication piece was a little clumsy the last time I tried it.
### The Problem
Unfortunately I ran into a problem immediately after the deployment completed:
![403 error from SSC](20211105_ssc_403.png)
Instead of being redirected to the vIDM authentication screen, I get a 403 Forbidden error.
I used SSH to log in to the SSC appliance as `root`, and I found this in the `/var/log/raas/raas` log file:
```
2021-11-05 18:37:47,705 [var.lib.raas.unpack._MEIV8zDs3.raas.mods.vra.params ][ERROR :252 ][Webserver:6170] SSL Exception - https://vra.lab.bowdre.net/csp/gateway/am/api/auth/discovery may be using a self-signed certificate HTTPSConnectionPool(host='vra.lab.bowdre.net', port=443): Max retries exceeded with url: /csp/gateway/am/api/auth/discovery?username=service_type&state=aHR0cHM6Ly9zc2MubGFiLmJvd2RyZS5uZXQvaWRlbnRpdHkvYXBpL2NvcmUvYXV0aG4vY3Nw&redirect_uri=https%3A%2F%2Fssc.lab.bowdre.net%2Fidentity%2Fapi%2Fcore%2Fauthn%2Fcsp&client_id=ssc-299XZv71So (Caused by SSLError(SSLCertVerificationError(1, '[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: self signed certificate in certificate chain (_ssl.c:1076)')))
2021-11-05 18:37:47,928 [tornado.application ][ERROR :1792][Webserver:6170] Uncaught exception GET /csp/gateway/am/api/loggedin/user/profile (192.168.1.100)
HTTPServerRequest(protocol='https', host='ssc.lab.bowdre.net', method='GET', uri='/csp/gateway/am/api/loggedin/user/profile', version='HTTP/1.1', remote_ip='192.168.1.100')
Traceback (most recent call last):
File "urllib3/connectionpool.py", line 706, in urlopen
File "urllib3/connectionpool.py", line 382, in _make_request
File "urllib3/connectionpool.py", line 1010, in _validate_conn
File "urllib3/connection.py", line 421, in connect
File "urllib3/util/ssl_.py", line 429, in ssl_wrap_socket
File "urllib3/util/ssl_.py", line 472, in _ssl_wrap_socket_impl
File "ssl.py", line 423, in wrap_socket
File "ssl.py", line 870, in _create
File "ssl.py", line 1139, in do_handshake
ssl.SSLCertVerificationError: [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: self signed certificate in certificate chain (_ssl.c:1076)
```
Further, attempting to pull down that URL with `curl` also failed:
```sh
root@ssc [ ~ ]# curl https://vra.lab.bowdre.net/csp/gateway/am/api/auth/discovery
curl: (60) SSL certificate problem: self signed certificate in certificate chain
More details here: https://curl.se/docs/sslcerts.html
curl failed to verify the legitimacy of the server and therefore could not
establish a secure connection to it. To learn more about this situation and
how to fix it, please visit the web page mentioned above.
```
In my homelab, I am indeed using self-signed certificates. I also encountered the same issue in my lab at work, though, and I'm using certs issued by our enterprise CA there. I had run into a similar problem with previous versions of SSC, but the [quick-and-dirty workaround to disable certificate verification](https://communities.vmware.com/t5/VMware-vRealize-Discussions/SaltStack-Config-Integration-show-Blank-Page/td-p/2863973) doesn't seem to work anymore.
### The Solution
Clearly I needed to import either the vRA system's certificate (for my homelab) or the certificate chain for my enterprise CA (for my work environment) into SSC's certificate store so that it will trust vRA. But how?
I fumbled around for a bit and managed to get the required certs added to the system certificate store so that my `curl` test would succeed, but trying to access the SSC web UI still gave me a big middle finger. I eventually found [this documentation](https://docs.vmware.com/en/VMware-vRealize-Automation-SaltStack-Config/8.6/install-configure-saltstack-config/GUID-21A87CE2-8184-4F41-B71B-0FCBB93F21FC.html#troubleshooting-saltstack-config-environments-with-vrealize-automation-that-use-selfsigned-certificates-3) which describes how to configure SSC to work with self-signed certs, and it held the missing detail of how to tell the SaltStack Returner-as-a-Service (RaaS) component that it should use that system certificate store.
So here's what I did to get things working in my homelab:
1. Point a browser to my vRA instance, click on the certificate error to view the certificate details, and then export the _CA_ certificate to a local file. (For a self-signed cert issued by LCM, this will likely be called something like `Automatically generated one-off CA authority for vRA`.)
![Exporting the self-signed CA cert](20211105_export_selfsigned_ca.png)
2. Open the file in a text editor, and copy the contents into a new file on the SSC appliance. I used `~/vra.crt`.
3. Append the certificate to the end of the system `ca-bundle.crt`:
```sh
cat <vra.crt >> /etc/pki/tls/certs/ca-bundle.crt
```
4. Test that I can now `curl` from vRA without a certificate error:
```sh
root@ssc [ ~ ]# curl https://vra.lab.bowdre.net/csp/gateway/am/api/auth/discovery
{"timestamp":1636139143260,"type":"CLIENT_ERROR","status":"400 BAD_REQUEST","error":"Bad Request","serverMessage":"400 BAD_REQUEST \"Required String parameter 'state' is not present\""}
```
5. Edit `/usr/lib/systemd/system/raas.service` to update the service definition so it will look to the `ca-bundle.crt` file by adding
```
Environment=REQUESTS_CA_BUNDLE=/etc/pki/tls/certs/ca-bundle.crt
```
above the `ExecStart` line:
```sh
root@ssc [ ~ ]# cat /usr/lib/systemd/system/raas.service
[Unit]
Description=The SaltStack Enterprise API Server
After=network.target
[Service]
Type=simple
User=raas
Group=raas
# to be able to bind port < 1024
AmbientCapabilities=CAP_NET_BIND_SERVICE
NoNewPrivileges=yes
RestrictAddressFamilies=AF_INET AF_INET6 AF_UNIX AF_NETLINK
PermissionsStartOnly=true
ExecStartPre=/bin/sh -c 'systemctl set-environment FIPS_MODE=$(/opt/vmware/bin/ovfenv -q --key fips-mode)'
ExecStartPre=/bin/sh -c 'systemctl set-environment NODE_TYPE=$(/opt/vmware/bin/ovfenv -q --key node-type)'
Environment=REQUESTS_CA_BUNDLE=/etc/pki/tls/certs/ca-bundle.crt
ExecStart=/usr/bin/raas
TimeoutStopSec=90
[Install]
WantedBy=multi-user.target
```
6. Stop and restart the `raas` service:
```sh
systemctl daemon-reload
systemctl stop raas
systemctl start raas
```
7. And then try to visit the SSC URL again. This time, it redirects successfully to vIDM:
![Successful vIDM redirect](20211105_vidm_login.png)
8. Log in and get salty:
![Get salty!](20211105_get_salty.png)
The steps for doing this at work with an enterprise CA were pretty similar, with just slightly-different steps 1 and 2:
1. Access the enterprise CA and download the CA chain, which came in `.p7b` format.
2. Use `openssl` to extract the individual certificates:
```sh
openssl pkcs7 -inform PEM -outform PEM -in enterprise-ca-chain.p7b -print_certs > enterprise-ca-chain.pem
```
Copy it to the SSC appliance, and then pick up with Step 3 above.
I'm eager to dive deeper with SSC and figure out how best to leverage it with vRA. I'll let you know if/when I figure out any cool tricks!
In the meantime, maybe my struggles today can help you get past similar hurdles in your SSC deployments.

Binary file not shown.

After

Width:  |  Height:  |  Size: 21 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 79 KiB

Binary file not shown.

After

Width:  |  Height:  |  Size: 109 KiB

Some files were not shown because too many files have changed in this diff Show more