[Linux-HA] usage of send_arp

Horms horms at verge.net.au
Sun Oct 2 21:29:39 MDT 2005


On Wed, Sep 14, 2005 at 10:18:08AM -0600, Alan Robertson wrote:
> Horms wrote:
> >On Wed, Sep 14, 2005 at 10:52:21AM +0800, ankh.tan wrote:
> >>BTW, whether send_arp supports VLAN interface, for example, eth0.1, etc.
> >>
> >>>On Tue, Sep 13, 2005 at 03:15:50PM +0800, ankh.tan wrote:
> >>>>For some reason, I need to use send_arp command in my scripts under
> >>>>HA enviroment. I want to know what meaning each parameter of the
> >>>>command is. any document or usage comment?
> >>>./heartbeat/libnet_util/send_arp
> >>>usage: send_arp: sends out custom ARP packet. packetfactory.net
> >>>       usage: send_arp [-i repeatinterval-ms] [-r repeatcount] [-p
> >>>pidfile] device src_ip_addr src_hw_addr broadcast_ip_addr netmask If
> >>>src_hw_addr is "auto" then the address of device will be used
> >>>
> >>>(I have no idea what the stray "packetfactory.net" is there for)
> >>>
> >>>Here is a breif description of what this means.
> >>>
> >>>
> >>>repeatinterval-ms: time in milliseconds between sending arp packets
> >>>
> >>>  Actually, that is a lie. For each arp requested,
> >>>  two are sent, an arp request, and an arp reply.
> >>>  This is becuse some systems ignore one or the other,
> >>>  and this gives the greates chance of success.
> >>>
> >>>  Each time an arp is sent, if another arp will be sent
> >>>  the code sleeps for half of repeatinterval-ms
> >>>
> >>>repeatcount: how many pairs of arp packets to send. See above for
> >>>  why pairs are sent.
> >>>
> >>>pidfile: pid file to use
> >>>
> >>>device:  interace to use
> >>>
> >>>src_ip_addr: source ip address
> >>>
> >>>src_hw_addr: source hardware address. If "auto" then the
> >>>            address of the device (interface will be used)
> >>>
> >>>broadcast_ip_addr: ignored
> >>>
> >>>netmask: ignored
> >
> >Incidently, are there any obections to adding the above text
> >to the help message that send_arp prints out?
> 
> No.  2.0.1 should go out today.  Then you can put this in (after the 
> code freeze).

I have committed this just now.

-- 
Horms



More information about the Linux-HA mailing list