[lvs-users] LVS - NAT and kernel panic

To: " users mailing list." <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Subject: [lvs-users] LVS - NAT and kernel panic
From: "pierrick grasland" <pierrick.grasland@xxxxxxxxx>
Date: Thu, 20 Mar 2008 18:04:07 +0100

I have been testing my lvs director (with ipvsadm and debian etch, kernel
2.6.18) for a basic SIP message (options), in order to find the capacity of
my box.

So, I have configure my LVS as in my previous message on round robin, and
setup a SIPp test (a testing tool for SIP).

*director:~# ipvsadm
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
UDP rr
  ->              Masq    1      0          0
TCP rr
  ->              Masq    1      0          0

I have only one realserver (because my client and my realserver are
overpowered compare to my director). I use old hardware (2-3 years old) for
my director (based around an athlon XP 1700+), which cannot compete with
core2duo and Athlon 3500+ in term of computing power.

But, when I reach something around 2000 messages/seconds, my director send
me a message about uncatch interruptions, and make a kernel panic.

When I restart it, it show me the following :

Begin:Running /scripts/init-bottom ...
run-init: /sbin/init: Permission denied
Kernel panic - not syncin: Attempted to kill init!

With a liveCD, I can browse my disk, but I have nothing in /etc
(input/output error), and I can't find anything useful on my logs (they have
stop to record when the kernel panic occurred)

So, did someone encounter before this type of problem ?

And, if you need some data, I can reproduce this pretty easily (I already
done it 2 times, I can go for more)

Grasland Pierrick

<Prev in Thread] Current Thread [Next in Thread>