RedHat Trouble netdump

Netdump Server and Client setup

netdump server: collard.alf.cpqcorp.net (16.113.5.139)
netdump client: bubba.alf.cpqcorp.net (16.113.5.104)

Setup Netdump Server:

1. Check to see if the netdump-server is installed:
#rpm –qa | grep –i netdump-server

If it is not already installed, install the netdump-server package.
#rpm –ivh netdump-server-0.6.11-3.i386.rpm

2. Set a password for the netdump user.
# passwd netdump
Changing password for user netdump
New password:
Retype new password:
passwd: all authentication tokens updated successfully

3. Optional: Copy the scripts files.
#cp /usr/share/doc/netdump-server-0.6.11/example_scripts/* /var/crash/scripts/

4. Configure and start the service.
# chkconfig netdump-server on
# service netdump-server start
Starting netdump server: [ OK ]

Setup Netdump client:

1. Check to see if the netdump-0.6.11-3 is installed.
#rpm –qa | grep netdump

If it not already installed, install the netdump client package
# rpm –ivh netdump-0.6.11-3.i386.rpm

2. Edit the file /etc/sysconfig/netdump.
Add a line NETDUMPADDR=x.x.x.x in which you specify the address of the netdump server.
NETDUMPADDR=16.113.5.139 (address of the server)

3. Get netdump init script to send a dynamic random key to the server.

# service netdump propagate
The authenticity of host ‘16.113.5.139 (16.113.5.139)’ can’t be established.
RSA key fingerprint is e8:67:5e:d3:a8:19:14:04:c4:9f:c3:74:ab:fb:c3:c8.
Are you sure you want to continue connecting (yes/no)? yes
Warning: Permanently added ‘16.113.5.139’ (RSA) to the list of known hosts.
netdump@16.113.5.139’s password:

(The password that was used for the netdump user on the server)

4. Configure and start the service

# chkconfig netdump on
# service netdump start
initializing netdump [ OK ]

To test the setup

On the client, this only works on the supported Ethernet drivers. As of now, the supported drivers are 3c59x, eepro100, e100, e1000, tlan, and tulip. RedHat suggests to check the /var/log/messages to see if you have a supported driver.

• Enable the magic sysrq key
# sysctl -w kernel/sysrq=1
• Enable panic_on_oops
# sysctl -w kernel/panic_on_oops=1
• Check to see if netlog is working
# echo h > /proc/sysrq-trigger
( A directory with the client’s IP address should be created on server’s /var/crash)
• Sync all mounted file systems
# echo s > /proc/sysrq-trigger
• Crash the system
# echo c > /proc/sysrq-trigger

On the netdump server’s 16.113.5.104-2005-03-15-21:15 directory two files should be created: log and vmcore.

To have the sysrq enabled on reboot:

#vi /etc/sysctl.conf
kernel.sysrq=1
kernel.panic_on_oops=1

Netdump whitepaper