failed to connect: dial unix /var/run/libvirt/libvirt-sock: connect: no such file or directory

[Alice Huston]

I'm getting the below error while running kubitect apply, it looks like its related to this issue in terraform-provider-libvirt that started after the 0.7.4 release.

│ Error: failed to connect: dial unix /var/run/libvirt/libvirt-sock: connect: no such file or directory
│
│   with provider["registry.terraform.io/dmacvicar/libvirt"].devops-host,
│   on main.tf line 25, in provider "libvirt":
│   25: provider "libvirt" {
│
╵
┌
│ Error: terraform plan failed: exit status 1

Here are the masked connection details I'm using in my kubitect.yml file

 connection:
      type: remote
      user: devops
      ip: 192.168.xxx.xxx
      ssh:
        port: xxx
        keyfile: "/home/devops/.ssh/id_ed25519"

In the meantime, if anyone else is experiencing this, I was able to fix the issue by making the below change in kubitect 3.2.2 and then building the repo. I know that may not be a long term solution, but I'm doing a lot of creating and destroying of my cluster while prototyping the final design and this was blocking that.

diff --git a/embed/terraform/versions.tf b/embed/terraform/versions.tf
index 6ee579c..7d2a1bd 100644
--- a/embed/terraform/versions.tf
+++ b/embed/terraform/versions.tf
@@ -8,7 +8,7 @@ terraform {
   required_providers {
     libvirt = {
       source  = "dmacvicar/libvirt"
-      version = "~> 0.7.1"
+      version = "0.7.1"
     }
   }
 }
(END)

[Din Mušić]

Hi @ahuston-0,

Thanks for pointing out the issue. Kubitect should use a specific version of libvirt terraform provider (as it is using a specific version of Terraform) to avoid such issues.

Could you open a PR with this change? I will make sure to include it in 3.2.3 and 3.3.0.

[Alice Huston]

Sure! will need a few hours but I can put something together.