vultr issue

I've had a small FreeBSD system hosted on Vultr for about 8 years. It is currently running 15.1-RELEASE. It is using ufs.

On the 18th Vultr had a maintenance window to update KVM to address a "critical Linux KVM vulnerability". They didn't indicate which issue or what version of KVM would be in place.

Since then, the server has been stuck in a boot loop. From the console I can see the boot menu come up, it loads the kernel, it begins booting from disk. It spins for 2 or 3 seconds and then restarts. To me that suggests that an incompatibility has been introduced to what ever version of KVM they have gone too.

I've opened a ticket of course but have any of you been having a similar experience or have any pointers?
 
First, we don't support vultr here. We are only FreeBSD users.

Seems the problem comes from a patch concerning "critical Linux KVM vulnerability".
We are again outside FreeBSD.

That being said, it can interest some people here.
 
I am quite aware this forum is for FreeBSD. I am asking other FreeBSD users if they have experienced a similar issue. I'm pretty sure that's very clear in the post.

At any rate, that matter has been resolved. KVM's latest version requires FreeBSD specific settings to run it (FreeBSD) as a client and the provider has made the required changes. Hopefully, if any other FreeBSD user comes across the issue with this major vendor they will know there is a fix. Sadly they did not provide any specific information.

Thank you so much for your eagerness to... well, help works since I'm feeling charitable.
 
I've had a small FreeBSD system hosted on Vultr for about 8 years. It is currently running 15.1-RELEASE. It is using ufs.

On the 18th Vultr had a maintenance window to update KVM to address a "critical Linux KVM vulnerability". They didn't indicate which issue or what version of KVM would be in place.

Since then, the server has been stuck in a boot loop. From the console I can see the boot menu come up, it loads the kernel, it begins booting from disk. It spins for 2 or 3 seconds and then restarts. To me that suggests that an incompatibility has been introduced to what ever version of KVM they have gone too.

I've opened a ticket of course but have any of you been having a similar experience or have any pointers?
Same issue - It was resolved by support in 15 minutes after I lodged a ticket. What wasn't resolved was the the UFS file system was somewhat hosed and needed three rounds of fsck. It finally booted but the system name was "amnesiac" and the root user was simply "0" and some command attempts resulted in "Who are you?" LOL. The ls binary was missing, but I remembered from an earlier UNIX time (Coherent, I think) that echo * did the same thing. Later I remembered that there was the rescue directory.

I manually typed in the commands to get the network up. The system was running FreeBSD 14.4p7, but I used freebsd-upgrade to upgrade to 14.4 setting "--currently-running to 13". On reboot things were better. The network started. The rc.conf was read. I just had to restore some missing binaries and libraries to get sendmail, some milters and apache running again.

There was a previous Vultr forced reboot in October last year after which the system suffered no ill effects. This time Vultr had to "update the XML definitions for your instance" to stop the boot loop. No idea why this time was different.
 
Back
Top