canonical / canonical/multipass
Configuration for network interface changed
- Dominant language
- C++
- Stars
- 9.2k
- Forks
- 828
- Avg merge
- 3d 1h
- Merged PRs (30d)
- 44
Description
**Describe the bug**
I am using multipass for microk8s and from time to time the k8s cluster becomes unreachable and the logs show connectivity/networking errors.
After a laptop restart, I noticed that many `multipass` commands (`list`, `start`, `stop`) became quite slow.
```
time multipass list -vvvv
[2024-08-24T00:10:46.913] [info] [microk8s-vm] SSH session disconnected
[2024-08-24T00:10:46.914] [debug] [microk8s-vm] Caching new SSH session
[2024-08-24T00:11:06.915] [debug] [microk8s-vm] Error getting extra IP addresses: ssh connection failed: 'Timeout connecting to 10.211.57.4'
Name State IPv4 Image
microk8s-vm Running 10.211.57.4 Ubuntu 22.04 LTS
multipass list -vvvv 0,02s user 0,01s system 0% cpu 34,785 total
```
Digging a little I noticed that, on the host interfaces, I don't have anything matching the IP address of the instance `10.211.57.4`, but I have this one instead:
```
bridge100: flags=8a63 mtu 1500
options=3
ether 5e:e9:1e:0b:44:64
inet 10.211.58.1 netmask 0xffffff00 broadcast 10.211.58.255
```
I noticed by some manual test that `10.211.58.3` responds to the ping and is actually my `microk8s-vm` and it is working as expected if I use that IP address.
**To Reproduce**
Unfortunately, I don't know... I noticed the different interface after a machine restart.
**Expected behavior**
The configured interface stays the same unless the user specifies it.
**Logs**
[multipassd.log](https://github.com/user-attachments/files/16735655/multipassd.log)
**Additional info**
- OS: 14.6.1 (23G93) - Apple M2 Pro
- `multipass version`
```
multipass 1.14.0+mac
multipassd 1.14.0+mac
```
- `multipass info`: `info failed: qemu-system-aarch64: Failed to load snapshot: Invalid argument`
- `multipass get local.driver`: `qemu`
Contributor guide
Research direction
Start with the attached multipassd.log and the reported `multipass list -vvvv` output, then inspect how the macOS host interface and the QEMU driver are selected after restart. Done means the configured interface remains stable and instance commands reach the VM without the reported timeout.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp, macos
- Domain
- networking, operating-systems
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 40/100