RE: Wtf? puzzling lock-ups
On 11-Aug-99 Jim wrote:
> Hi. As the subject suggests, I get strange freeze-ups on my machine. Yes,
> it's definitely frozen... I can't ping it from my lan, and nothing
> responds... no keybd, no mouse, no nothing.
>
> This occurs once every couple of weeks. So far I have not been able to
> get more than 21 days of uptime, since installing Debian 2.1 on this
> machine in mid-April of this year.
This sound VERY familiar! I have had the samething, but I could ping my box
for some crazy reason and ipmasq still worked. I couldn't telnet in, though.
> The funny thing that I started to notice the last time it locked up is
> that the system clock jumps when it occurs. I had been noticing funny
> syslogs in the sense that I would see normal log entries occurring at,
> let's say, 10:00, then it would lock up and I'd reboot and then I'd see
> the end of the syslog contain new records that were from 9:45... 15 mins
> EARLIER. It took me a while to put 2 and 2 together, but I can safely say
> now that each time the machine freezes, the clocks either skips ahead or
> behind anywhere from 15-25 minutes, and I have to re-set the time after
> rebooting.
I figured out that was why my clock keeps going off... yesterday morning my
system froze, and now I see that it is 30 minutes AHEAD.
> Another interesting thing is this part of my /var/log/syslog file:
>
> Aug 11 01:12:55 my_machine pppd[14060]: sent [LCP EchoReq id=0x60
> magic=0x6e5674f0]
> Aug 11 01:12:55 my_machine pppd[14060]: rcvd [LCP EchoRep id=0x60
> magic=0x4f4b52b8]
> Aug 11 01:13:25 my_machine pppd[14060]: sent [LCP EchoReq id=0x61
> magic=0x6e5674f0]
> Aug 11 01:13:25 my_machine pppd[14060]: rcvd [LCP EchoRep id=0x61
> magic=0x4f4b52b8]
> ^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
> ^@
> ^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
> ^@
> ^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
> ^@
> ^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@
> ^@
>
> The next log entry from pppd would have occurred 30 seconds after the last
> one, except this is where my machine froze... and apparently whatever
> process was writing to the syslog file at the time got screwed and just
> dumped garbage into it.
I just checked my syslog, and I get the normal pppd entries:
Aug 10 08:04:46 gateway pppd[21520]: rcvd [LCP EchoRep id=0x7 magic=0x2938d119]
Aug 10 08:05:14 gateway pppd[21520]: sent [LCP EchoReq id=0x8 magic=0xa606a722]
Aug 10 08:05:16 gateway pppd[21520]: rcvd [LCP EchoRep id=0x8 magic=0x2938d119]
And then this when it froze and I had to hit the reset button:
Aug 10 08:09:10 gateway syslogd 1.3-3#32: restart.
And the rest of the normal boot procedure:
Aug 10 08:09:10 gateway kernel: klogd 1.3-3#32, log source = /proc/kmsg started.
Aug 10 08:09:10 gateway kernel: Inspecting /boot/System.map-2.2.10
> It MAY only be occurring while X is running, although I can't really
> verify that since X is always running. The reason I say that though is
> because the last 3 times this happened, it was while I was dragging
> windows around or picking items from drop-down lists in Netscape (4.51
> Comm).
I use X all the time too. It has happened when Netscape is not running, and
with different versions of Netscape.
> I am running XFree86 3.3.3.1-2 (from netgod.net/x), although this also
> happened with the stock slink XFree86. I'm on a 2.0.37 kernel... but
> again this was an issue in 2.0.36 as well. Window manager is Window Maker
> 0.53. Hardware is a P133, Asus P/I P55T2P4 mobo, 64 MB RAM, NE2000 NIC,
> Hercules Dynamite 128/Video card (2 MB).
I am using the latest potato version of X, 2.2.10, WindowMaker 0.6x, AMD
K6-350, Atrend MB, 128 MB SDRAM, PCI NE2000 NIC, Creative SBlive!, 8MB AGP ATI
3D Rage Pro video card. Oh, and I am always running a couple of Setiathome
clients as well.
> Has anyone got any ideas about where I might look for a solution to this?
> Unfortunately I cannot make this happen at will, it seems to be quite
> unpredictable.
Same here.
> Also, any (free) software I can get that would be able to perform some
> thorough tests on my RAM and CPU? Perhaps it's one of those ubiquitous
> "hardware issues."
Maybe the ol' kernel recompile?
> Thanks for any assistance provided.
See
http://www.debian.org/Lists-Archives/debian-user-9908/msg00293.html
for somebody's idea to use "Magic SysRq" in the kernel.
Summary: I have a totally different hardware, and different kernel/x/other
software versions, but we are still experiencing the same symptoms.
Regards,
---------------------------
Wim Kerkhoff
wim@netmaster.ca
www.canadianhomes.net/wim
ICQ: 23284586
Reply to: