-
Notifications
You must be signed in to change notification settings - Fork 88
Add troubleshooting and workaround details of VM IP address missing #863
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Conversation
…lapping Signed-off-by: Renuka Devi Rajendran <[email protected]>
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Technically LGTM, just one nit. Thank you.
Check the ip address on the pod interface of the VM created using the following command which matches the ipv6 address updated in the Virtual Machine Instance. | ||
|
||
```shell | ||
$ kubectl exec -it -n <PODNAMESPACE> <PODNAME> -- /bin/bash | ||
$ ip a | ||
``` |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It'd be more user-friendly if we could instruct them to look at which pod interface.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM, thanks.
BTW: if issue also exists on v1.5.x, please copy the changes to versioned_docs/version-v1.5/
branch.
|
||
Qemu guest agent is responsible for notifying the interface details and guest os info from the vm guest os to the Virtual Machine Instance which is displayed by the Harvester UI. | ||
During the issue, qemu guest agent is able to get the network information and update the status to the Virtual Machine Instance but sometimes domain Info holding the network interface details becomes empty during the updates. | ||
This problem is due to an upstream kubevirt issue occuring spcifically on alma linux 9 and rocky linux 9 guest os version. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
spcifically -> specifically
``` | ||
|
||
### Workaround | ||
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If the VM's IP is still useable for e.g. ssh/curl, please add an info section to cease user's potential worry.
|
||
### Issue Description | ||
|
||
The IP address field of the Virtual Machines screen on the Harvester UI flaps for a newly created Virtual Machine. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Can we clarify what "flapping" means? Also, other than it being an UI issue, what operational impact does this have on the virtual machine? Will the IP address eventually re-appear as described here? More importantly, does that mean the virtual machine is fully unreachable? Would restarting the virtual machine help? User can still access it via the serial console for debugging?
|
||
The IP address field of the Virtual Machines screen on the Harvester UI flaps for a newly created Virtual Machine. | ||
|
||
## Issue Analysis |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
## Issue Analysis | |
### Issue Analysis |
|
||
https://github.com/kubevirt/kubevirt/issues/12698 | ||
|
||
## Virtual Machine IP Address not displayed when VM Pod Interface has IPv6 address |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It feels like these two issues have very similar symptoms. In the first case, the IP address "flapped", which I assume it eventually re-appears. For the second case, it seems like the effect is more permanent?
I suggest combining the two issues into one section titled "Missing Virtual Machine IP Address In Harvester UI". Then provide details on the different debugging steps and a set of possible mitigation/recovery actions. Depending on the actual cause, some recovery actions may be more relevant than others. The goal is to help user confirm whether if the IP is flapping, caused by IPv6, or other guest OS level issues. E.g., with Ubuntu Noble cloud image, I see the IP address disappearing issue more often when the guest runs with only 1 CPU on my KVM setup. Should the user try restarting the virtual machine, how does the user check if the virtual machine is still reachable? If not reachable, then the virtual machine is officially down. Can the user still access it via serial console, to run dmesg
etc.
Problem:
Users are unaware of why IP address is missing in Harvester UI sometimes
Solution:
Troubleshooting and workaround for ip address missing and flapping in Harvester UI
Related Issue(s):
harvester/harvester#3990
harvester/harvester#6955
Test plan:
To reproduce:
case 1:
1.Create a VM using alma linux 9 or rocky linux 9
2.Wait for VM to move into running state
3.Login to the guest os and check ip a (this will display a v4 address acquired from DHCP or configure it statically)
4.Check if Harvester UI displays the ipv4 address in the UI (flaps intermittently)
case 2:
1.Create a VM using any guest OS
2.Wait for VM to move into running state
3.Login to the guest os and check ip a (this will display a v4 address acquired from DHCP or configure it statically)
4.Login to the pod of the VM and check using "ip a".The pod will have v6 link local address obtained using slaac.
4.Check if Harvester UI displays the ipv4 address in the UI (displays nothing)
Additional documentation or context