UPDATE - emulators/virtualbox-ose version 6.1.18 committed to head

Wow, we are matching upstream. For such a complex bit of software this is really impressive!

I use this a fair amount and would be cool to test. I have fairly old packages installed (dependencies for the older VBox version installed) but lets see if the port can build against what I have.
 
_martin said:
Note: for FreeBSD 13 and newer.
Click to expand...
You are mistaken, please see uname (haven't tested functionality yet):

virtualbox.jpg
 
Hmm. Can you test it too ? Preferably with the VMs that use host-only interfaces too. RL is keeping me way too busy but I'll try to find some time to test it myself on 12.2 too.
If this is true then I stand corrected.
 
My lasts tests I did on 12.2 were with 6.1.16. With 6.1.18 I was able to fully build it. New VM created with the vbox6 seems to be working ok, at least my first tests show so.
Problem is when you try to start VMs that were created with 5.x version:
Code: 
# VBoxManage startvm box01 --type headless
Waiting for VM "box01" to power on...
VBoxManage: error: The virtual machine 'box01' has terminated unexpectedly during startup with exit code 1 (0x1)
VBoxManage: error: Details: code NS_ERROR_FAILURE (0x80004005), component MachineWrap, interface IMachine
#
This happens also if you build a new VM and attach the old disk to it.

EDIT: if I remove the modules ( /usr/local/etc/rc.d/vboxnet stop) and then start it again I can start the new VM with the old disk. I can't start VM that was built in 5.x.
Then I imported other VMs. Did a reboot and now I can start both new and old VMs. I don't like these heisenbugs. But for now yeah, 6.1.18 seems to be running under 12.2.

EDIT2: I stressed it a bit more and I got kernel panic out of it (same trace as mentioned already in the bug report).

EDIT3: I did some additional testing. It seems when the DHCP is off for that interface it does work. There's still bug in the kernel module but this way it's not easily triggered by the VirtualBox commands.
 
_martin said:
Hmm. Can you test it too ? Preferably with the VMs that use host-only interfaces too. RL is keeping me way too busy but I'll try to find some time to test it myself on 12.2 too.
If this is true then I stand corrected.
Click to expand...
RL = Rocky Linux?
 
I'm just wondering as I'm too lazy to test this. Has anyone tried this with Windows 10 and found it's quicker? I was using VB with a Windows 10 VM and it was painfully slow. I then switched to bhyve, and while still slow, it's much better. On a CentOS laptop, Windows 10 in VB runs quite well--the laptop is much newer and on an nvme drive while the Virtual machine on my tower is on a spinning drive.
 
scottro said:
I'm just wondering as I'm too lazy to test this. Has anyone tried this with Windows 10 and found it's quicker? I was using VB with a Windows 10 VM and it was painfully slow. I then switched to bhyve, and while still slow, it's much better. On a CentOS laptop, Windows 10 in VB runs quite well--the laptop is much newer and on an nvme drive while the Virtual machine on my tower is on a spinning drive.
Click to expand...
I never tried Windows 10 with the older port. So I did a install of Windows 10 with this new version to test it out. It took forever on the "Installing Updates" steps. Beyond that I don't notice it to be any slower than my other guest machine: Windows XP, Solus, or FreeBSD (with MATE desktop).

One quality of life improvement I've seen is that sound on my FreeBSD guest is no longer choppy. However I am still having mouse pointer accuracy glitches on the FreeBSD guest. Also you can't install the extension pack to get USB 2.0 (same issue as before). This is not to discredit the hard work it took to get here. These issues are with VirtualBox itself.

Code: 
Failed to install the Extension Pack Oracle_VM_VirtualBox_Extension_Pack-6.1.18.vbox-extpack.

Failed to locate the main module ('VBoxPuelMain').

Result Code: NS_ERROR_FAILURE (0x80004005)
Component: ExtPackManagerWrap
Interface: IExtPackManager {70401eef-c8e9-466b-9660-45cb3e9979e4}
 
scottro So you're saying newer notebook had better performance ? I've only Win7 on current 5.x port and performance is ok-ish (very,very subjective). I had to toggle few things in VM setup to make it work though. I'm not sure if virtio drivers work there but usually it's a good start when troubleshooting performance.
 
_martin Yes, and it was VirtualBox on CentOS, as opposed to VirtualBox on FreeBSD. In my case, I need Windows on very rare occasion for stuff at work, and found bhyve gave me acceptable performance. As you said about your case, very subjective. I haven't tried a Windows VB on the new lapop with FreeBSD, which I probably should, to compare, but for my use case, it's not really worth it, as one has to do X, Y, and Z to avoid creating an MS account, and so on.
Also, this is Windows 10, not 7.
 
scottro Yeah, avoiding MS account is pain but still doable. I'm afraid you'll have to do the tests yourself as only you'll be able to compare it to what you're used to.
 
_martin said:
scottro Yeah, avoiding MS account is pain but still doable.
Click to expand...
The Windows OBE won't prompt for an MS Account if it doesn't have Internet access at first boot.

Anyway, back to the topic. I noticed a slight issue with my Windows 10 VM where it locked up. This is what was in the log before I had to reboot the host.

Code: 
00:01:49.638071 AIOMgr: Host limits number of active IO requests to 256. Expect a performance impact.
00:02:00.625782 AIOMgr: I/O manager 0x000008127d18e0 encountered a critical error (rc=VERR_FILE_AIO_NO_REQUEST) during operation. Falling back to failsafe mode. Expect reduced performance
00:02:00.625827 AIOMgr: Error happened in /wrkdirs/usr/ports/emulators/virtualbox-ose/work/VirtualBox-6.1.18/src/VBox/VMM/VMMR3/PDMAsyncCompletionFileNormal.cpp:(1677){int pdmacFileAioMgrNormal(RTTHREAD, void *)}
00:02:00.625839 AIOMgr: Please contact the product vendor
00:02:29.425376 AHCI#0: Port 0 reset
00:02:29.426438 VD#0: Cancelling all active requests
00:02:29.426456 VD#0: Request{0x00000812a09ac0}:
00:02:29.426457     Type=WRITE State=ACTIVE Id=0x8 SubmitTs=12259943 {28513} Flags=0x2
00:02:29.426459     Offset=3741368320 Size=131072 Left=131072 BufSize=131072
00:02:29.426481 VD#0: Request{0x00000812a09c80}:
00:02:29.426482     Type=WRITE State=ACTIVE Id=0x1 SubmitTs=12259826 {28630} Flags=0x2
00:02:29.426483     Offset=4097142784 Size=65536 Left=65536 BufSize=65536

EDIT: never-mind, I failed to read the pkg-message.
 
You must log in or register to reply here.
Back
Top