VM aborts/freezes

[mksmr]

Good evening everybody, I'm running this combo of machines:

Host:
HP DL380 G6, 24 GB RAM
OS: 4.9.0-8-amd64 #1 SMP Debian 4.9.110-3+deb9u4 (2018-08-21) x86_64 GNU/Linux

Guest:
8 CPUs, PAE/NX active, PIIX3 chipset, IO-APIC active
16 GB RAM
Debian 9.5
accessed through RDP, <hostip>:3390

The machine is running well but has hung several times., the screen freezing, forcing me to reboot. On the virtualbox control panel it was marked with a red "no entry" sign and the word "abgebrochen" (aborted).

I admit the VBox.log is rather confusing to me. I've been running a number of guests on virtualbox for quite a while and never had to dig deeper because they simply used to run without any issues. This is the first that's giving me a bit of trouble.

I wondered if anybody could give me a hint where to begin and what to look for in VBox.log. At first glance I couldn't see anything obvious there.

TIA
Matthias

[andyp73]

I wondered if anybody could give me a hint where to begin and what to look for in VBox.log. At first glance I couldn't see anything obvious there.

You might find it quicker to zip the VBox.log file and post it here.

If you want to search for yourself the obvious things to look for are any references to guru meditations, that the host still has at least one physical CPU core available for its own use and that the allocated memory (system RAM and VRAM) are less than the available memory reported near the top of the file.

-Andy.

[mksmr]

What I find particularly annoying about the log is that it's not logging the actual time, but what looks like the seconds the guest has been running after boot. I wondered if there's a way to change that. Unless I'm missing something, logging doesn't seem to be very well documented.

[mksmr]

After another hang I've found this obvious thing in VBox.Log:

Code: Select all

220:54:20.567193 Display::handleDisplayResize: uScreenId=0 pvVRAM=00007f393ccbe000 w=1920 h=1080 bpp=0 cbLine=0x1E00 flags=0x5
224:46:04.089257 VMMDev: vmmDevHeartbeatFlatlinedTimer: Guest seems to be unresponsive. Last heartbeat received 5 seconds ago
224:46:04.494669 VMMDev: GuestHeartBeat: Guest is alive (gone 6 395 792 383 ns)
224:46:09.756825 VMMDev: vmmDevHeartbeatFlatlinedTimer: Guest seems to be unresponsive. Last heartbeat received 5 seconds ago
(End of file)

I've attached the gziped file.

Matthias

[andyp73]

VirtualBox VM 5.2.12 r122591 linux.amd64 (May  9 2018 15:00:04) release log

Life has moved on. The current released version of VirtualBox (as of writing) is 5.2.22. You might want to upgrade to see if that helps any.

00:00:00.033858 Host RAM: 24101MB (23.5GB) total, 16225MB (15.8GB) available
00:00:00.223100   RamSize           <integer> = 0x0000000466100000 (18 892 193 792, 18 017 MB, 17.5 GB)
00:00:00.224205   VRamSize         <integer> = 0x0000000010000000 (268 435 456, 256 MB)

You have allocated memory to the guest that you don't have available on the host. Your options are: a) allocate less memory to the guest, b) close some applications on the host, or c) buy some more memory for the host. In the first instance, try reducing it to something like 8.0GB.

00:00:00.223083   NumCPUs           <integer> = 0x0000000000000008 (8)
00:00:00.801648 CPUM: Physical host cores: 8

You have assigned all of your physical CPU cores (the thing VirtualBox cares about) to the guest. This may lead to resource starvation on the host. You are going to have to reduce the guest CPU count to give the host a bit of breathing space. Try setting it to 4.

-Andy.

[mksmr]

Resource exhaustion on the host system then? I used the settings allowed by the GUI. Does the log give more hints? I must admit I'm a noob to reading virtualbox logs. I've reduced the guest resources to 6CPUs and 16384 MB now.

[mksmr]

The guest machine still crashes, reproduceable when I am running a ClamTk-Scan. I cannot tell at which point, though, because the filesystem I'm scanning is quite big. So far, no "guru meditation" messages in the logs. Logs attached.

[mksmr]

...and VBox.log.3