On 21 July 2014, Liviu Daia <[email protected]> wrote:
> On 12 July 2014, Stuart Henderson <[email protected]> wrote:
> > trace ->
> >
> > stopped at uao_reference+0x88: movq %rcx,0x8(%rax)
> >
> > uao_reference at ..+0x88
> > uao_set_swslot at ..+0x55
> > uvmpd_scan_inactive at ..+0x681
> > uvmpd_scan at ..+0x23c
> > uvm_pageout at ..+0x5b
> >
> > active process is pagedaemon
> >
> > screenshots at
> > https://drive.google.com/folderview?id=0B8t-sinTZPnucERodGdCcTc2Mmc&usp=drive_web
> >
> > machine was untarring files at the time (reading from softdep on
> > softraid, writing to a tmpfs /tmp, which <since this> i have now
> > replaced with mfs).
>
> Possibly related: I get a panic upgrading a test virtual machine
> to today's snapshot. Screenshot:
>
> https://drive.google.com/file/d/0B-zUWP5Vwmx1cVZVZlg3NDdYV2c/edit?usp=sharing
[...]
After some more testing, this seems more likely to be related to
a bug in the latest version of VirtualBox, rather than a problem in
OpenBSD. I don't have definitive evidence either way, but (1) I can
easily reproduce the problem with VirtualBox 4.3.14, basically any burst
of disk activity (tar, CVS update, etc.) triggers a panic, and (2) I
downgraded VirtualBox to 4.3.12 and couldn't trigger the problem while
playing around for a day. FWIW.
Regards,
Liviu Daia
