Strange Network issues with new install

[timothy carroll]

I installed hp-ux v23 on 5 rx2600 integrity machines - I gave them all addresses in
a 170.154.182.7x range and I could get to them fine from my machine.

Then the were located to racks - 2 of them have ip addresses in the
170.154.223.X range and they work fine and I can get to them

2 of them were put into another rack and there addresses are 170.154.224.X and I cannot ssh to them from my machine - I can ssh between the 2 machines if I am on the console -

My question is this:

If I put my laptop or a linux box on the 224.X network - I can ssh to them - but not the HP-UX boxes...

Has anyone come across this issue....

[Michal Kapalka (mikap)]

hi,

what about the default GW ??? and dns ( name resolving )

mikap

[timothy carroll]

I will post my netconf for both boxes

for the 224 network -

INTERFACE_NAME[0]="lan0"
IP_ADDRESS[0]="170.154.224.46"
SUBNET_MASK[0]="255.255.255.192"
BROADCAST_ADDRESS[0]=""
INTERFACE_STATE[0]=up
DHCP_ENABLE[0]=0
INTERFACE_MODULES[0]=""

ROUTE_DESTINATION[0]="default"
ROUTE_MASK[0]=""
ROUTE_GATEWAY[0]="170.154.224.1"
ROUTE_COUNT[0]="1"
ROUTE_ARGS[0]=""


for the 223.x boxes...

INTERFACE_NAME[0]="lan2"
IP_ADDRESS[0]="170.154.223.54"
SUBNET_MASK[0]="255.255.255.192"
BROADCAST_ADDRESS[0]=""
INTERFACE_STATE[0]=up
DHCP_ENABLE[0]=0
INTERFACE_MODULES[0]=""

ROUTE_DESTINATION[0]="default"
ROUTE_MASK[0]=""
ROUTE_GATEWAY[0]="170.154.223.1"
ROUTE_COUNT[0]="1"
ROUTE_ARGS[0]=""

[timothy carroll]

There is no dns for these boxes - only entry in my resolv.conf is a domain

the nsswitch.conf file has files for all entries....

If a do an nslookup - it resolves through the hosts file

[timothy carroll]

here is my netstat -rn for the 223. network

-bash-4.0# netstat -rn
Routing tables
Destination Gateway Flags Refs Interface Pmtu
127.0.0.1 127.0.0.1 UH 0 lo0 4136
170.154.223.54 170.154.223.54 UH 0 lan2 4136
170.154.223.0 170.154.223.54 U 2 lan2 1500
127.0.0.0 127.0.0.1 U 0 lo0 0
default 170.154.223.1 UG 0 lan2 0

224. is same except all the 223's are 224

[Michal Kapalka (mikap)]

Hi,

the configuration is correct,

maybe there is a miss configuration oon the network level ( Switch - gw ), try to ping the other GW, for example from 224/26 == the IP adress of the GW from 223.

mikap

[timothy carroll]

I cannot ping, ssh, etc from the 223 to the 224 network - I can howevere ping and ssh from the 224 to the other 224's - it is not machine specific, as we tried changing other hp-ux boxes to the 224 network and the same issue happens...

The strange thing is, if I hook the 224's to the 223 switch they work fine - (provided I add 223 addresses and gateway)

Also remember, a linux box works fine on the 224 and even my windows box

[V. Nyga]

Hi,

what's about changing
SUBNET_MASK[0]="255.255.255.192"

to
SUBNET_MASK[0]="255.255.0.0"?

because you want to change to another subnet with a different third number ...

Volkmar

[Michal Kapalka (mikap)]

hi,

strange, but did you try to switch the HP-UX box to your functional LINUX LAN from 224 network ??? maybe there is something wrong on the first 64 IP form 224.

mikap

[timothy carroll]

the linux and windows box are plugged into the same switch as the hp-ux boxes - I gave them all 224.X addressses and same netwmask and gateway -

I can ssh to them from 223 and ping them - I cannot ping or ssh to the 224's

Its not the switch, since I tried another switch on the 224 network....

[Michal Kapalka (mikap)]

Hi,

ask the network administrator if there is some rules between Vlan224 and Vlan 223, maybe there are some restriction.

mikap

[Jozef_Novak]

Hi Timothy,

I guess that your subnet mask is definitely wrong. You are using the third octet for both network and host identification, while the netmask suggests that the entire third octet is still in network portion.

J.

[timothy carroll]

so you think I should change the subnet mask to something like 255.255.0.0

[timothy carroll]

ok - so I tried changing subnet mask - now here is an extra gotcha - when i restart network

/sbin/init.d/net start

I can get to it for about 2-3 minutes and then nothing...

I just ssh'd from 223 and about 1 minute later I could not...

[timothy carroll]

-bash-4.0# netstat -rn
Routing tables
Destination Gateway Flags Refs Interface Pmtu
127.0.0.1 127.0.0.1 UH 0 lo0 4136
170.154.224.49 170.154.224.49 UH 0 lan3 4136
170.154.0.0 170.154.224.49 U 2 lan3 1500
127.0.0.0 127.0.0.1 U 0 lo0 0
default 170.154.224.1 UG 0 lan3 0
-bash-4.0# ifconfig lan3
lan3: flags=1843
inet 170.154.224.49 netmask ffff0000 broadcast 170.154.255.255

[Duncan Edmonstone]

> I can get to it for about 2-3 minutes and then nothing...


usually means you have a gateway that doesn't respond to ICMP echo requests (pings)

try disabling dead gateway detection:

# ndd -set /dev/ip ip_ire_gw_probe 0

you can make the change permanent by adding an entry to /etc/rc.config.d/nddconf

HTH

Duncan

[timothy carroll]

made the change - restarted network - no go....

It was not only a ping that would stop, but if I had ssh'd to the 224, the session would hang