Cannot ping package from different subnet

[Rgomes]

Hi All,

My Problem:

We cannot ping or connect to a vitual IP (10.16.1.3) of a ServiceGuard package from a PC host (10.32.15.28/12) in a different subnet. Two HP9000 rp7400 servers (HP-UX11i) are configured under the MC ServiceGuard. The PC host is using Windows OS.

Primary Server IP (lan0): 10.16.1.2 (sm: 255.240.0.0)
Failover Server IP (lan0): 10.16.2.2 (sm: 255.240.0.0)
vedas_package virtual IP (lan0:1): 10.16.1.3 (sm: 255.240.0.0)
PC client IP : 10.32.15.28 (sm: 255.240.0.0)


folowing is the route table
/#netstat -vr
Routing tables
Dest/Netmask Gateway Flags Refs Interface Pmtu
localhost/255.255.255.255 localhost UH 0 lo0 4136
vedas_pkg/255.255.255.255 vedas_pkg UH 0 lan0:1 4136
hps1_dc/255.255.255.255 hps1_dc UH 0 lan0 4136
hps1_dc-hb/255.255.255.255 hps1_dc-hb UH 0 lan2 4136
192.168.36.0/255.255.255.0 hps1_dc-hb U 2 lan2 1500
10.16.0.0/255.240.0.0 hps1_dc U 3 lan0 1500
10.16.0.0/255.240.0.0 vedas_pkg U 3 lan0:1 1500
10.32.0.0/255.240.0.0 10.16.8.1 UG 0 lan0 0
127.0.0.0/255.0.0.0 localhost U 0 lo0 0
default/* 10.16.8.1 UG 0 lan0 0
/#
/#hostname
hps1_dc
/#

Is there any way/route add entry for this vedas_pkg, so we can ping it from another subnet?

Thanks all

Richard

[Rajeev Shukla]

Its looks alright from server side.
Is this happening from all PCs' in that subnet?
Check the Default gateway on those PC if that is set to 10.16.8.1(thats what looks like your gateway for both subnets).

Rajeev

[Ron Kinner]

It does appear that the problem is probably on the other end or perhaps the router is blocking pings. On a window box you can add a route with:

route add 10.16.0.0 mask 255.240.0.0 10.32.x.y -p

where 10.32.x.y is the router which knows how to get to the 10.16.0.0 network. Presumably this is the same router which also has IP address 10.16.8.1 but you should use an address on the same 10.32 LAN.

You might also want to run a traceroute from the window box to see where it fails:

tracert -d 10.16.1.3

Also check that the mask on both router interfaces is correct and that the router is not blocking ICMP echo or ICMP echo-reply.

Finally if the PC has two NICs make sure that you have only assigned a default gateway to one NIC. Having two default gateways just confuses it.

Ron

[Sridhar Bhaskarla]

Hi Richard,

10.32.0.0/255.240.0.0 10.16.8.1 UG 0 lan0 0


The above entry is not required as 10.32.0.0 network is different from that of the local network and you already configured the default route wiht the same gateway. But this shouldn't cause any issues.

Check the routing table on your PC and make sure that a default router is defined there too. You will need to have a router (or a switch router) to route the packets between your PC and the hp system as both are in different subnets. Then add the router to your PC using the "route add" command.

Or if possible, configure your PC with an IP address anywhere in 10.16.0.0 - 10.32.0.0 range and you don't need a router.

-Sri

[U.SivaKumar_2]

Hi,

After starting the Cluster with Virtual IP address , clear the ARP cache of the router or reboot the router once.

Eg:-

Cisco router

clear arp

regards,
U.SivaKumar

[Ravi_8]

Hi.

from your windows OS PC
C:> route add

[Jim Keeble]

Hi Richard,

Assuming you can ping the primary ip 10.16.1.2 from the remote subnet, it's not likely that the problem is on the server.

See if you can ping vedas_pkg from one of the hosts on the 10.16.0.0 subnet.

If so, then try to ping the vedas_pkg from the router. If it fails, check the arp table on the router to see if the MAC address of vedas_pkg is the same as hps1_dc.

You shouldn't need the 10.32.0.0 route either way, since it uses the default gateway.

[Rgomes]

Hi Rajeev,

"...Check the Default gateway on those PC if that is set to 10.16.8.1(thats what looks like your gateway for both subnets)."

MY GATEWAY IS ALRIGHT, WHICH IS 10.32.21.1( ROUTER E0/1), AND I CAN PING 2xHP SERVERS LAN0 IP.


Hi Ron,

AFTER, 1)route add 10.16.0.0 mask 255.240.0.0 10.32.x.y -p

D:\>tracert -d 10.16.1.2

Tracing route to 10.16.1.2 over a maximum of 30 hops

1 <1 ms <1 ms <1 ms 10.32.21.1
2 7 ms <1 ms <1 ms 10.16.1.2

Trace complete.

D:\>tracert -d 10.16.1.3

Tracing route to 10.16.1.3 over a maximum of 30 hops

1 <1 ms <1 ms <1 ms 10.32.21.1
2 * * * Request timed out.
3 * * * Request timed out.
4 * * * Request timed out.
5 ^C
D:\>

2)Also check that the mask on both router interfaces is correct and that the router is not blocking ICMP echo or ICMP echo-reply.

ROUTER IS NOT BLOCKING ANY ICMP ECHO/REPLY.

Hi Sridhar,

YES, I KNOW THAT WILL MAKE MY LIFE MUCH EASIER, BUT WE NEED IT IN DIFFERENT SUBNET. I AM VERY VERY CONFUSED THAT, ALTHOUGH 2 HP SERVERS IN DIFFERENT SUBNET I CAN PING THEM (LAN0), BUT CANNOT PING THE PACKAGE IP.


Hi Sivakumar,

I WILL TRY THAT.

hI Jim,

1) See if you can ping vedas_pkg from one of the hosts on the 10.16.0.0 subnet

YES I CAN PING vedas_pkg FROM 10.16.0.0 SUBNET FROM ANY PC.

2)If so, then try to ping the vedas_pkg from the router. If it fails, check the arp table on the router to see if the MAC address of vedas_pkg is the same as hps1_dc.

ROUTR FAILS TO PING vedas_pkg( 10.16.1.3), SEE BELOW:

AS_DC#ping 10.16.1.2

Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 10.16.1.2, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 1/1/4 ms
AS_DC#ping 10.16.2.2

Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 10.16.2.2, timeout is 2 seconds:
!!!!!
Success rate is 100 percent (5/5), round-trip min/avg/max = 1/1/1 ms
AS_DC#
AS_DC#ping 10.16.1.3

Type escape sequence to abort.
Sending 5, 100-byte ICMP Echos to 10.16.1.3, timeout is 2 seconds:
.....
Success rate is 0 percent (0/5)
AS_DC#


SO, WHERE AM I?



RICHARD

[avsrini]

Hi Richard,
Try adding this.

route add net 10.16.0.0 netmask 255.240.0.0 10.16.1.3 1

This will add a route for 10.16.0.0 network on interface 10.16.1.3, your package's ip address.
After this you should be able to ping from your pc.

Give it a try and let us know.

Srini.