Re: lookup on memory shortage

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Fri, Oct 03, 2008 at 10:02:50AM -0500, Eric Haszlakiewicz wrote:
> hmm... I recently threw some extra ram into one of my machines, and to test it
> out a bit wrote a program that allocates and touches memory in a loop.  One
> of the times I ran it, it got killed because the system ran out of memory
> (which is what I expected) and the system recovered.  The other times, almost
> everything hung except that I was able to hit ^C (on a text wscons console)
> and it recovered.  Perhaps your hang is similar?  Does running this program
> reproduce your problem?

It does reproduce the issue. Just like you, the first run the program got
killed, on the second run the box did hang. Here's what top showed 1s
before the hang:
load averages:  2.29,  1.17,  0.97;               up 0+07:05:15        20:16:07
40 processes: 3 runnable, 36 sleeping, 1 on CPU
CPU states:  1.0% user,  0.0% nice, 99.0% system,  0.0% interrupt,  0.0% idle
Memory: 301M Act, 147M Inact, 12M Wired, 7176K Exec, 16M File, 16K Free
Swap: 256M Total, 256M Used, 4K Free

  PID USERNAME PRI NICE   SIZE   RES STATE      TIME   WCPU    CPU COMMAND
25038 bouyer    43    0   748K  160K RUN        0:03  4.20%  0.59% hang
  367 bouyer    85    0   824K  664K select     0:04  0.35%  0.05% sshd
    0 root      96    0     0K   12M nfsiod     2:55  0.00%  0.00% [system]
 6898 bouyer    43    0   756K  732K CPU        0:13  0.00%  0.00% top
  388 bouyer    85    0   764K  420K select     0:04  0.00%  0.00% screen-4.0.3
    1 root      85    0   748K    4K wait       0:04  0.00%  0.00% <init>
 4211 root      85    0   748K    4K wait       0:02  0.00%  0.00% <pbulk-build
 9457 root      85    0   128K  316K RUN        0:00  0.00%  0.00% sh
  360 bouyer    43    0   756K   12K RUN        0:00  0.00%  0.00% screen-4.0.3

(it's now 20:21:52 and the display didn't get updated).
On console:
Out of memory allocating ksiginfo for pid 219
+++++
Stopped in pid 0.23 (system) at netbsd:breakpoint+0x4:  popl    %ebp
db> tr
breakpoint(ffffff00,80,ca797f68,5,1,6,ca797fa8,c03aa6ff,ca77268c,c05bc005) at 
netbsd:breakpoint+0x4
xencons_tty_input(ca77268c,c05bc005,1,c03a3b9b,3b9aca00,0,6,0,4,2) at 
netbsd:xencons_tty_input+0xa6
xencons_handler(ca77268c,ca79abf0,0,6,c02de99e,4,0,0,c03a1b85,2) at 
netbsd:xencons_handler+0x5f
evtchn_do_event(2,ca79abf0,ca79aba8,0,fatal page fault in supervisor mode
trap type 6 code 0 eip c038e6d1 cs 9 eflags 10246 cr2 ca798000 ilevel 8
kernel: supervisor trap page fault, code=0

Faulted in DDB; continuing...
db>

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--