Abort diagnosis please

Discussions about using non Windows and Linux guests such as FreeBSD, DOS, OS/2, OpenBSD, etc.
Post Reply
grahamperrin
Posts: 57
Joined: 27. Apr 2008, 15:19

Abort diagnosis please

Post by grahamperrin »

Please, can anyone tell the reason for the abort in this case?

It was the first run of a new guest.

Attempted live boot from hello-0.5.0_0E43-FreeBSD-12.2-amd64.iso (helloSystem 0.5.0 (0E43), this build based on non-patched FreeBSD 12.2-RELEASE).

Unfortunately I was away from the screen when the abort occurred. The tail of the log:

Code: Select all

…
00:00:45.982781 NAT: Link up
00:00:47.435745 PGMR3PhysAllocateLargePage: allocating large pages takes too long (last attempt 1293 ms; nr of timeouts 1); DISABLE
00:00:47.538133 Audio Mixer: Setting recording source of sink '[Recording] Line In' to '[Line-In] OSS'
00:00:47.548467 Audio Mixer: Setting recording source of sink '[Recording] Line In' to '<None>'
00:00:47.549958 Audio Mixer: Setting recording source of sink '[Recording] Line In' to '[Line-In] OSS'
00:00:47.550031 Audio Mixer: Setting recording source of sink '[Recording] Line In' to '<None>'
00:00:47.550929 Audio Mixer: Setting recording source of sink '[Recording] Line In' to '[Line-In] OSS'
00:00:47.554613 PIIX3 ATA: Ctl#0: RESET, DevSel=0 AIOIf=0 CmdIf0=0x24 (-1 usec ago) CmdIf1=0x00 (-1 usec ago)
00:00:47.564706 PIIX3 ATA: Ctl#0: finished processing RESET
00:00:47.664977 PIIX3 ATA: Ctl#1: RESET, DevSel=0 AIOIf=0 CmdIf0=0xa0 (-1 usec ago) CmdIf1=0x00 (-1 usec ago)
00:00:47.675035 PIIX3 ATA: Ctl#1: finished processing RESET
Any 'loose' diagnosis might help; then I should be able to tell whether it was a known issue with helloSystem. Thanks.
Attachments
helloSystem 0.5.0 (0E43)-2021-03-28-12-30-35.log
VirtualBox log
(70.36 KiB) Downloaded 27 times
fth0
Volunteer
Posts: 5661
Joined: 14. Feb 2019, 03:06
Primary OS: Mac OS X other
VBox Version: PUEL
Guest OSses: Linux, Windows 10, ...
Location: Germany

Re: Abort diagnosis please

Post by fth0 »

Although the helloSystem%200.5.0%20%280E43%29-2021-03-28-12-30-35.log file ends "in the middle" and doesn't show any fatal error, there are three aspects I'd question:

1. The first log message timestamp was logged after nearly 21 seconds (typical values are far below 10 seconds). What happened during the first 20 seconds?

2. The "allocating large pages takes too long" log message indicates either severe host memory fragmentation (no contiguous 2 MB block of memory) or a 1293 ms delay for any other unknown reason. What could make it take so long?

3. In the previous years, I've only seen one or two users reporting VirtualBox problems and using zfs, and they also had timing issues. This has no statistical significance, of course.

Finally, a general question: Since the log file doesn't show any indication of an abort, what do you call an abort? Please provide more details.
grahamperrin
Posts: 57
Joined: 27. Apr 2008, 15:19

Re: Abort diagnosis please

Post by grahamperrin »

Thanks for replying.

1. Good question. I don't know the answer, but if I recall correctly there was a comparable delay at the head of the subsequent log file, for a boot that succeeded.

2. I don't know whether it's relevant, but sometimes (rarely) there's extreme slowness during a stage of live boot that involves uzip. In the photograph at https://github.com/helloSystem/ISO/issu ... -801714692 there were transfer rates of 29 and 14 MB/s; in the second screenshot at https://github.com/grahamperrin/ISO-hel ... m/issues/1 the rate was 889 kB/S and there was hogging of the CPU.

My best guess is that the kernel panicked i.e. https://bugs.freebsd.org/bugzilla/show_ ... ?id=254318 however on most (all?) past occasions I think a panic in this context is followed by continued running of the guest; the boot process (re)starts.

Aborted was the word shown in the main VirtualBox window when I found the guest unexpectedly not running.
fth0
Volunteer
Posts: 5661
Joined: 14. Feb 2019, 03:06
Primary OS: Mac OS X other
VBox Version: PUEL
Guest OSses: Linux, Windows 10, ...
Location: Germany

Re: Abort diagnosis please

Post by fth0 »

FWIW, there are additional global log files in $HOME/.config/VirtualBox/ you could look at.

I skimmed your threads and some basic information on the helloSystem GitHub pages, and it seems to be in some experimental state. If I download the latest release (hello-0.4.0_0D26-FreeBSD-12.1-amd64.iso) and try to run it in a VirtualBox VM, would you rather expect me to encounter problems or not?
grahamperrin
Posts: 57
Joined: 27. Apr 2008, 15:19

helloSystem 0.4.0 (0D26)

Post by grahamperrin »

Much of helloSystem is experimental, in this context 0.4.0 (0D26) is considered stable. I should not expect a kernel panic or any other issue at boot time.
Post Reply