[Help] Strange Linux server no response problem, fixed by hitting ENTER at console!

[Help] Strange Linux server no response problem, fixed by hitting ENTER at console!

Post by Charle » Thu, 22 Jul 2004 16:50:11


Dear all,

Recently one of our servers randomly stop responding, but it does not hang.

When the problem happens, all the network services stop responding, but
the server is still running. Then many processes are queued up and cause the
avg no. of process running sky to over 100 in 10min! If I was connecting an
ssh to the server, it just stop response but the connection is still there.
ping also has response.

The most strange thing is, when I press an ENTER key at the console, the
problem solved instantly and everything back to normal. Even my ssh connection
will have response again.

Our server's config:

Hardware: Dual P3-700Mhz, 640M RAM, 2 x 10G IDE HDD running Software RAID-1
OS: RedHat Linux 8.0
kernel: kernel-2.4.18-14smp

Special config:
- freeswan-1.99_x509_0.9.15_2.4.18_14 installed

Do you know what happens and kindly give me any suggestions? Thanks!


Best Regards,
Charles
 
 
 

[Help] Strange Linux server no response problem, fixed by hitting ENTER at console!

Post by robertharv » Fri, 23 Jul 2004 08:40:47

"Charles" < XXXX@XXXXX.COM > wrote

Hmm.
I think we need a bit more info
- are you running in text mode (init 3) or a graphical desktop (init
5)?
- when you say the console, do you mean that - or do you mean at any
prompt, or at any time with the X display running?
- WHat prompts/messages appear on your console before you press
'enter'?
- what does tail -100 /var/log/messages tell us?

 
 
 

[Help] Strange Linux server no response problem, fixed by hitting ENTER at console!

Post by Charle » Sat, 24 Jul 2004 11:25:24

Hi Robert,

It is a text mode installation with no X-Window, the console I hitting ENTER is the mingetty login prompt with no user login.

It reported different things in 2 different incidents:
1) init: Sending process the TERM signal
2) net: run message suppressed

I cannot find any errors in /var/log/messages, but I found some strange error message in /var/log/kern.log:
Jul 23 10:08:27 server kernel: sending pkt_too_big (len[1500] pmtu[1443]) to self
Jul 23 10:09:51 server kernel: sending pkt_too_big (len[1500] pmtu[1443]) to self
Jul 23 10:14:16 server kernel: sending pkt_too_big (len[1500] pmtu[1443]) to self
Jul 23 10:15:00 server kernel: Dead loop on netdevice ipsec0, fix it urgently!
Jul 23 10:15:00 server kernel: Dead loop on netdevice ipsec0, fix it urgently!
Jul 23 10:15:33 server kernel: sending pkt_too_big (len[1500] pmtu[1443]) to self
Jul 23 10:20:12 server kernel: sending pkt_too_big (len[1500] pmtu[1443]) to self

I did some searches from the net but cannot find any clues related to our server's problem:


Best Regards,
Charles