17:28:00 <natefoo> Hey all, I have a bhyve VM I'm trying to resize the root partition of using a gparted iso. At first I was getting only a black screen from the VNC console, but was able to fix that by switching the bootrom to UEFI. However, after X starts, the mouse is unusable.
17:28:38 <bahamat> natefoo: Yeah, there's a known issue that affects the mouse driver.
17:29:08 <natefoo> d'oh
17:29:18 <natefoo> Any workarounds?
17:30:06 <bahamat> In general the vnc console is only intended for auxiliary access when the instance is inaccessible via the network.
17:30:26 <bahamat> So something like X display forwarding will still work.
17:30:52 <natefoo> I don't think that's an option for the gparted live CD?
17:31:19 <bahamat> Can you use gparted from the command line?
17:33:11 <natefoo> Oh, yeah, well, I can presumably figure out (non-g)parted.
17:36:45 <natefoo> Well, I broke something, possibly trying to start bhyve by hand: vm_open: No such file or directory
17:37:03 <natefoo> prctl: 61a99139-0148-4da0-8d6c-903c1554e7d3: No controllable process found in task, project, or zone.
17:37:38 <bahamat> Yeah, that's not going to work.
17:38:08 <bahamat> bhyve runs with extremely limited permissions. It's part of our "double hulled" virtualization technology.
17:38:51 <natefoo> This is now when trying to start it the normal way from vmadm (after trying to start it by hand)
17:40:54 <bahamat> Unfortunately not knowing exactly what you did to get into this situation, it's difficult to say how to remedy it.
17:41:42 <bahamat> Even if you destroyed the partition table on the zvol you should still be able to get the bhyve process to start and say there's no bootable media.
17:41:59 <bahamat> I hope you had a back up?
17:43:36 <natefoo> So it was booting ok (without the mouse) to the CD image just after updating bootrom and bhyve_extra_args, afaik the only thing I did after that was to attempt to run the bhyve command (grabbed from ps) from the command line
17:43:47 <natefoo> I made a snapshot of disk0 before I started
17:44:02 <bahamat> maybe your extra args are causing the problem?
17:45:38 <natefoo> # ls -l /zones/61a99139-0148-4da0-8d6c-903c1554e7d3/root/usr/
17:45:38 <natefoo> total 0
17:45:48 <natefoo> No idea how I did that...
17:45:52 <natefoo> Well, hooray for snapshots
17:48:12 <natefoo> Oh, I guess that would be mounted by vmadmd at start?
18:01:46 <bahamat> Yeah, it's empty if the zone isn't booted.
18:02:27 <natefoo> Weird, ok. What does the ENOENT up there from vm_open refer to?
18:02:34 <natefoo> i.e. what is missing?
18:03:52 <bahamat> I'm not sure.
18:04:19 <bahamat> There might be some more detailed information in the zone_bh log
18:04:44 <bahamat> That'll be /var/log/zone_bh.<uuid> on the compute node.
18:06:58 <natefoo> I have quite a few zone_bh files in there but not one for this uuid and most are from 2019
18:53:19 <natefoo> Ok, dunno, but I got it back by creating a new empty VM and zfs cloning the snapshot over as the disk
19:21:14 <natefoo> And successfully resized with cli parted
19:21:21 <natefoo> Thanks bahamat 
19:48:19 <bahamat> natefoo: That's good to hear.