traceroute to cluster alias IP

[Francois Rabe]

Can anybody tell me if you can do a traceroute to the default cluster alias. If I do it, it returns nothing, but when I issue the following command:

traceroute -s xxxx yyyy

where xxxx is the source ip and yyyy the destination, it works fine.

traceroute xxxx does not work.

[Simon Hargrave]

This should work fine, the cluster alias is simply an IP address piggy backed to one of your cluster nodes. To all intents and purposes it's exactly the same as any other server with this IP address.

However I find it strange that you say:

---

traceroute -s xxxx yyyy

where xxxx is the source ip and yyyy the destination, it works fine.

traceroute xxxx does not work.

---

From the first bit I see you tracing yyyy as the destination, however the second bit you say you can't traceroute xxxx, which surely is the source?

Perhaps if you posted more information it may be clearer, eg: -

ip address of machine you are tracing from
real ip's of cluster nodes
alias ip address

Also the output of the working and the not working traceroute.

[Francois Rabe]

Sorry..this should read..

traceroute -s xxxx yyyy

where xxxx is the source ip and yyyy the destination, it works fine.

traceroute yyyy does not work.

[Steve Lewis]

This is because on HP-UX 11i if you specify the -s [source IP] it will use the routing table to correctly work out which interface to trace from.

It would also work if you use -i and specify the PPA of the lan interface on that subnet.

Without either of those parameters traceroute just tries the first interface on the lanscan list. If there is no route from that interface to your target IP, then it won't come back with anything.

The behaviour of traceroute changed from 11.00 to 11i.