Re: current kernel + userland just resets

["John D. Baker" <jdbaker%mylinuxisp.com@localhost>]

On Thu, 29 Sep 2011, matthew green wrote:

> this should be fixed now, assuming it is the same problem:
>
> http://mail-index.netbsd.org/source-changes/2011/09/28/msg027660.html
>
> it will have triggered under memory-pressure conditions.  the easy
> test case was building qemu in a smaller VM.  i noticed that at
> various points of the qemu build, cc1 grew to *over* 1GiB, and a
> gas process had grown to over 400MiB.  so.. qemu is a good test of
> straining the system.  this was also PR#45361.  the symptoms could
> have been kind of anything since it was a memory corruption issue.

Possibly.  I don't think what I was seeing was this, exactly.  My
main observation was that the machine would reboot but be unresponsive
to all user input devices or network traffic (I caught it early
enough once to see that XDM had started, but it didn't respond
to mouse or keyboard and the hard drive light was on solid).

With an iteration of 5.99.56 installed, it successfully resumes
operation following a panic reboot.

(Now I need to figure out why it keeps mangling a directory thus
triggering a panic when the maintenance scripts run at 3:15 every
morning.  I forcibly 'fsck'ed the disk and dealt with the fallout,
but it did it again now also mangling the package database in the
bargain.)


--
|/"\ John D. Baker, KN5UKS               NetBSD     Darwin/MacOS X
|\ / jdbaker[snail]mylinuxisp[flyspeck]com    OpenBSD            FreeBSD
| X  No HTML/proprietary data in email.   BSD just sits there and works!
|/ \ GPGkeyID:  D703 4A7E 479F 63F8 D3F4  BD99 9572 8F23 E4AD 1645