Do you have a question? Post it now! No Registration Necessary. Now with pictures!
- Allan Palm
October 3, 2004, 4:05 pm
rate this thread
an server that
runs slackware on it and it is in my garage. it's only task is to share internet
So sometimes i check it from my laptop... like today. Some how it didn't want to
make the ssh
connection so i started pinging it... it had 1000 ms delay. i thought that maby
it was my wiring
problem.(i moved this server to garage some days ago and did wiring myself) but
then some seconds
later it didn't ping it at all. i went to garage and it seemed that box had
crashed. so i rebooted
server. it booted fine, then i tried to connect again, again same problem.. it
but when i started to check log files it crashed again... only ssh connection...
ping went to 1000 ms
then to unreachable again... so i shutted down everything, don't know why i did
everything and booted server up again. now everything was ok and i started to
check log files.
didn't find anything at all... didn't know what went wrong... then i found
something from my syslog.
i didn't know what it was, but it has Process sshd in it and if i thought what
had just happend i
came up with an idea that maby it's ssh problem.... so maby someone could help
me out, maby saying
that it's nothing or it is because i turned power off or something else... thank
sorry but i didn't find any other useful data :(
linux kernel 2.4.26
Oct 3 18:19:30 Traktor kernel: Unable to handle kernel paging request at
virtual address 080f4da0
Oct 3 18:19:30 Traktor kernel: printing eip:
Oct 3 18:19:30 Traktor kernel: c01379a2
Oct 3 18:19:30 Traktor kernel: *pde = 0f236067
Oct 3 18:19:30 Traktor kernel: *pte = 00000000
Oct 3 18:19:30 Traktor kernel: Oops: 0000
Oct 3 18:19:30 Traktor kernel: CPU: 0
Oct 3 18:19:30 Traktor kernel: EIP: 0010:[<c01379a2>] Not tainted
Oct 3 18:19:30 Traktor kernel: EFLAGS: 00010202
Oct 3 18:19:30 Traktor kernel: eax: 080f4d88 ebx: 00000001 ecx: 080f4d88
Oct 3 18:19:30 Traktor kernel: esi: cc9573c0 edi: cc9573c0 ebp: cc2b57a8
Oct 3 18:19:30 Traktor kernel: ds: 0018 es: 0018 ss: 0018
Oct 3 18:19:30 Traktor kernel: Process sshd (pid: 692, stackpage=c9b59000)
Oct 3 18:19:30 Traktor kernel: Stack: c0270a0c c9b59f64 cc9573c0 9a489320
c0270a43 cc9573c0 c032be98 cc9573c0
Oct 3 18:19:30 Traktor kernel: c0270b9c cc9573c0 00000000 00000030
00000000 c02929b3 cc9573c0 00000000
Oct 3 18:19:30 Traktor kernel: c9b59f64 00000030 00000000 c0119ef4
cc2b5790 c9b58000 00000001 00000000
Oct 3 18:19:30 Traktor kernel: Call Trace: [<c0270a0c>] [<c0270a43>]
[<c0270b9c>] [<c02929b3>] [<c0119ef4>]
Oct 3 18:19:30 Traktor kernel: [<c02aedb0>] [<c026d357>] [<c026d48c>]
Oct 3 18:19:30 Traktor kernel:
Oct 3 18:19:30 Traktor kernel: Code: 8b 40 18 f6 c4 40 75 0a ff 49 14 0f 94 c0
84 c0 75 01 c3 89
Oct 3 18:25:18 Traktor kernel: Linux version 2.4.26 (root@Traktor) (gcc version
3.3.4) #2 Sun Sep 26 23:22:06 EEST 2004
Re: kernel paging problem
A wiring problem can't cause such delays.
A packet passes through Ethernet or loses.
Your server died when you had pinged it?
Did you try only ping it, without ssh?
Kernel memory faults are a common symptom of a badly compiled Linux kernel.
E.g. I have it when built a Linux 2.2.26 on buggy GCC 3.2.2.
You can debug your kernel in a such way:
$ gdb /usr/src/linux-2.4.26/vmlinux
(gdb) disassemble 0xc01379a2
Then you can see the crashed kernel routine disassembled
and check its C source to get a clue what really happened.
I give path to kernel image in assumption that your kernel is a custom kernel.
Error in application process should not crash a kernel in normal circumstances,
so your problem seems a bit offtopic in comp.security.ssh.
Moreover, such crash log says virtually *nothing* about your sshd
except that sshd process existed at the time of the crash.
BTW, sshd process should always exist ;)
I added crosspost and followup
to more topical newsgroup news:comp.os.linux.misc ,
perhaps not sufficiantly topical, but _more_ topical.
Also, hardware problems may be responsible for these crashes.
Perhaps, you should have several crash logs to distinguish
hardware problems from software ones.
You should attach the console to your box (just insert a VGA card)
because it is more likely to obtain a crash report in the console
than kernel would be in time to store it on the HDD.
Try check your server's networking using simpler tools than ssh
but more sophisticated than just iptables routing on it;
iptables produce almost no system load.
Ping it, set up simple network services (echo, qotd, daytime ) there, test it.
Or better, just test you suspect box locally.
FYI a text of 102 columns width considered /de mauvais ton/ in Usenet.
You should try to keep your narrative text in 78 columns,
but this restriction is not applicable to logs.
qq~~~~\ [ úá IP âåú ãåîúõòù ]
/ /\ \ [ FAQ you ] € http://www.comtv.ru/~av95/chainik.html
\ /_/ /
\____/ Linux console notes http://www.comtv.ru/~av95/linux/console /
- » domain users not able to login w/ cygwin/openssh - domain admins works ok
- — Previous thread in » Secure Shell Forum
- » ssh on command line: force using a group size (prime size) of 1024 (and no...
- — Newest thread in » Secure Shell Forum