bhyve VM exited on signal 10

[janb14]

Yesterday I upgraded my Freenas 11 to the 11.2 Beta-RC1 after it got released to get the new UI and have IOcage integration. Everything is working as expected but my bhyve Vm that runs Pihole keeps on crashing after the upgrade and i didnt change anything on it. For the Record it didnt crash a single time before and im lost about the reason.

Code:

Jul 10 19:26:34 JansCloud kernel: pid 15240 (bhyve), uid 0: exited on signal 10
Jul 10 20:52:03 JansCloud kernel: pid 38653 (bhyve), uid 0: exited on signal 10
Jul 10 22:44:14 JansCloud kernel: pid 74956 (bhyve), uid 0: exited on signal 10
Jul 11 02:33:33 JansCloud kernel: pid 38235 (bhyve), uid 0: exited on signal 10

Thats what i get when i go : cat /var/log/messages | grep bhyve
and after i restart the VM every single time.
The VM running is ubuntu server on 17.10 with only a PiHole install.

 

[jclendineng]

Yep mine is doing the same thing, doesn't matter the vm. Weird...exiting on signal 10.

 

[jclendineng]

Im going to wipe and clean install and see if its just an upgrade bug

 

[jclendineng]

A couple clean installs later and it still shuts off with the same error so it's a bug for sure. Might want to report it. Tested beta1 and latest master as well as Ubuntu, and centos VMs and it doesn't matter what I do, eventually it shuts off with an error 10.

 

[ryanakata]

I have the same exact issue. Upgrade to Beta 1 went fine. VM keeps stopping and kills my network. The VM currently runs Pihole and Unifi for me. Moved off the Pihole for now and will move back to it once the VMs are fixed.

 

[dlavigne]

Someone please create a report at bugs.freenas.org and post the issue here so that those affected by it can follow its progress.

 

[jclendineng]

Fixed it i think. Intel e1000 driver does not work. Virtio see to fix the issue. Weird!

So try that and report back. I'll monitor my VMS today but it sat overnight without crashing so I'd say it is a bad Intel e1000 driver.

 

[ryanakata]

Fixed it i think. Intel e1000 driver does not work. Virtio see to fix the issue. Weird!

So try that and report back. I'll monitor my VMS today but it sat overnight without crashing so I'd say it is a bad Intel e1000 driver.

Mine won't even boot when switched to VirtIO. Creating a report now. Also noticed there is no generate mac option in the new GUI.

 

[jclendineng]

No, I just grab the auto generated mac from the box using ifconfig and place it in a static lease of my choosing on my gateway. You need to disable wait to connect under VNC as well.

Also, the device boots with e1000? Hmm. Shouldn't affect booting. Its a driver issue when the device has already booted, nothing that affects starting/stopping.

 

[jclendineng]

Can confirm that 100% fixes the issue. Should use virtio anyways, MUCH better performance on virtualized hardware.

 

[ryanakata]

No, I just grab the auto generated mac from the box using ifconfig and place it in a static lease of my choosing on my gateway. You need to disable wait to connect under VNC as well.

Also, the device boots with e1000? Hmm. Shouldn't affect booting. Its a driver issue when the device has already booted, nothing that affects starting/stopping.

Wait to connect is currently disabled and was on the 11.1 install as well. I created bug 37842. I'll just use Google DNS for now and reload a UniFi backup once it is stable again.

 

[jclendineng]

Do a clean install...other issues with upgrading that I fixed by clean install. Create a snapshot, install and re add the vm.

Id do a clean install for the VM anyways if its only dns...takes like 15 minutes to throw ubuntu 18.04 on a vm and install pihole. For that matter get a pi! Always should have physical dns/gateway/dhcp just in case. I use vms for all that but I have physical as a backup just in case.

 

[janb14]

Glad im not the only one. It gets even stranger. I just removed my vnc and added 512mb more ram just in case. Runs stable for over 24h now. Maybe we just need to change any setting and it rewrites the config files and that fixes the error?

 

[jclendineng]

I was getting spammed in the logs, just a ton of errors. A reinstall fixed it and logging is fixed :)

more /var/log/messages
Or use less

 

[ryanakata]

Do a clean install...other issues with upgrading that I fixed by clean install. Create a snapshot, install and re add the vm.

Id do a clean install for the VM anyways if its only dns...takes like 15 minutes to throw ubuntu 18.04 on a vm and install pihole. For that matter get a pi! Always should have physical dns/gateway/dhcp just in case. I use vms for all that but I have physical as a backup just in case.

I have a couple of Pis I can use as a backup and that's no worries but I do have Tautulli and UniFi running in the VM too.

I tried to create one from scratch and even it halts immediately after starting so for now I'm holding on until I see it get fixed in a nightly or RC1.

 

[jclendineng]

It works...we just need to figure it out

 

[jclendineng]

What error does it halt on?

 

[ryanakata]

What error does it halt on?

Same one as earlier. I was on 16.04 and tried the newer version as well.

Code:

kernel: pid 12428 (bhyve), uid 0: exited on signal 10

 

[jclendineng]

Thats a network error....I tested on 17.10 and 18.04, 16.04 is a bit old, 18 adds lots of driver fixes. Can you test a newer ubuntu?

 

[ryanakata]

Thats a network error....I tested on 17.10 and 18.04, 16.04 is a bit old, 18 adds lots of driver fixes. Can you test a newer ubuntu?

Tried to come in and install Debian, VNC comes in for a couple of seconds and then closes when the VM stops. Log message is sig 11 rather than 10 though. The VM will not start after that initial stop though.


kernel: pid 54444 (bhyve), uid 0: exited on signal 11