camrossi/akb

ACI Kubernetes Bootstrapper

akb's Issues

new proxy config breaks ha-proxy on secondary masters

prepare sandbox: set the apt mirror back to the default once the update is complete

Package the app as a container

vc_template_upload: Implement progress bar

This is for the branch vc_ova_upload for now. The upload percentage is already calculated but is not displayed on the page. We should add that

Previous button should re-populate the state

do not create k8s cluster not working

when using the do not create k8s cluster option the system breaks as it cant find the directory to copy the calico files from.
need to fix the dir structure.

VM: pre-execute terraform init

Make Dual Stack Optional

Currently the cluster is always a dual stack v4/v6 environment. This is not always needed and would be easier for the end use to have this as an optional feature.

router id: 1.1.1.1 and so on

loadInput() should be page aware

loadInput() override the Cluster BGP AS value in the cluster_network because the HTML ID is called local_as for both the l3out page and the cluster_network page.

For now I did a quick fix and changed the HTML page to use k8s_local_as

Pick the AAEP and create a Phy Dom and create the Port Group

Support deploying cluster if there is no internet access

Remove Node Config

Since moving to the one AS per cluster there is no more need to have multiple Node Config.
The AS can be set at the BGPConfiguration

replace apt mirror should not run if ubuntu_apt_mirror is empty.

Remove L3out name and pre-configure one

when you go to /cluster directly, UI breaks

this is due to the fact vc dict is not created and vc["vm_deploy"] does not exist.

K8s Cluster Hide it all under advanced

Document or cross link vkaci documentation.

bgp peer can't be established if VRF has "BD Enforcement Status" enabled

"BD Enforcement Status" prevent any passive bgp peer connection from being established.

VM: Add customizations from vCenter

IP, DNS, GW etc...

Support Creating Multiple Clusters

update k8s version automatically

l3out hide all under physical domain

K8s VM template

Currently the tool expects a VM template to be exisitng in the VCenter. This template will be used to clone VMs and install K8s. It would be nice if the app can embed the template image and in the workflow the user can request to create the VM template with the embedded image.

Generate worker join token: Should only be executed onces

Don't crash if the IPv4 or IPv6 address of the node is invalid

ipaddress.AddressValueError
ipaddress.AddressValueError: Empty octet not permitted in '192.168.0.'

Open PreReq as a new TAB

l3out page: Copy the IPv4 and IPv6 value to the Node Primary IP

For example if I do this:

I want the IPv4 and IPv6 value to be copied to the Node Primary IPv4 and IPv6

Webpage to scroll to the bottom out the text output

Destroy AKB user when cluster is destroyed

in the create and destroy page you use the config on file and not the "vc" variable

the vc variable is not always present in memory.
We should just load the HLC config we have on file

Make the installer VM Ubuntu

and add static DNS entries to reach the services like neo4j

Randomize AKB User and Password

handle http https or no protocol in the APIC URL

Add Option to download Calico Manifests

vCenter Details: Going back from calico_nodes looses Folder and Cluster name

Document better L3out and contract requirements

Enter keypress moving the application to the previous page

Create application releases

Automate packaging of the app as a VM

Create the VM automatically and upload to cloud storage

Package the app as standalone pip like install

Fix Formatting for submit button

For some reason the formatting is not working very well anymore

Also the pop up/helper instead of showing (i) shows [] on Chrome on Linux

If you need to test you can run the application from BM-01 in the lab @mntxca

Grey out IPv6 fields if no IPv6 subnet is provided

check_apiserver.sh should be created before configure keepalived StaticPOD

If the images are pre-pulled the StaticPOD comes up before the check_apiserver.sh is created and this leads to failures.
Just change the order of operation in ansible. i.e. first create the check_apiserver.sh

Gracefully handle vmware login failure

support access /cluster directly for advanced configs

This used to work but I broke it

Document subnet allocations

tracking ER to add and remove leaves from L3out

this issue is to track the capability of the tool to allow adding and removing leaf switches from L3out after the initial execution of the tool.
ideally if a cluster exists, we should not load the 'existing_cluster' but rather give option to user to delete or modify existing cluster. If user wants to modify, then he/she will be prompted with the option to change the leaves configured for the L3out.

git action: Check that ACI is receiving the expected rotues

packer: VM should have netaddr and sshpass installed