Netdump over bonded interface ?

Linux and Unix-Like Red Hat
1

Netdump over bonded interface ?
Hey there,..
I have heard from admins, that netdump usually fails over a bonded network interface.

I have written a multi-OS checkdump script , which checks if the respective crashdump utility on corresponding OS is correctly configured, space is available, etc etc etc.

For the netdump check, I wanted to add in a check to see if netdump is using bonded interface . How do I do that ? I can easily find out if the system has bonding enabled or if bond0 is up or down .
But I want to know if Netdump is going to use this bonded interface to send out the dump, or the usual interface ?

/etc/sysconfig/netdump doesnt give much of an idea. Does LOCALPORT setting have something to do with it ? It defaults to 6666.
I have minimal understanding about the networking side of this problem, could some one please tell me :

  1. How do I find out which network interface (outgoing) netdump is currently configured to use ? How to check if its bonded or not ??

  2. I also wanted to do a ethtool $NIC to check if the ethernet link is
    down or up, so this brings back to finding out which network interface is used by netdump , eth0 , eth1 , bond0 ??? How can I find this out.

Thanks,
Shrikant

2

Hello,

If you want to check if netdump is correctly configured and working?

By the client side, when you start netdump the return code doesn't work. I have a client that fails in start and return always exit code 0.

And what about if you check the server side? Each time you start a netdump client, a new log is generated like this:

[root@box]# cat /var/crash/172.19.113.119-2009-08-24-12:35/log
[...network console startup...]

. I think you could try to restart at same time all netdump servers by cron and test only one time in netdump server.