Static Routing

[Idham]

Hi,

If any expert can help.

My rx46460 (11i v2)server have 2 lan interfaces.
lan0 : 172.130.8.197 (production)
lan1 :192.169.1.240 (backup)

backup server ip :192.169.1.40

The purpose of this routing is to make the lan1 traffic dedicated for backup purposes and will not congest the production traffic.

I did run this command....

route add host 192.169.1.40 192.169.1.240

But somehow all the traffic(both lan) use the backup lan.So I need to disable the routing.

Currently the the backup use the production lan which reduce the throughput.

Please help.

[V. Nyga]

Hi,

check '/etc/rc.config.d/netconf' there should be both interfaces and it's also the right place for a permanent routing.

HTH
Volkmar

[sreekanthtm]

Hi,

Your Lan1 IP (192.169.1.240) and the backup server IP (192.169.1.40) is in the same subnet. So need not to add the route.
You can simply fire backup from the production server. It will use the 192.169.1.240 IP only. beacuse It cannot communicate to the other lan interface (Different Class IP is assigned to lan0).

So simply delete the route and you can achieve the target.

Rgds
Sreekanth

[Laurent Menase]

Hi
When asking such question you should include
a netstat -nrv

because like that like already said, lan0 and lan1 have not the same class,

so traffic to net 192.169 should go lan1
and to net 172.130 go through lan0

execept if the remote address ids in lan1 range, or that default gateway s in lan1 network.

[Idham]

Below is the output from netstat

camracd0:>netstat -rnv
Routing tables
Dest/Netmask Gateway Flags Refs Interface Pmtu
127.0.0.1/255.255.255.255 127.0.0.1 UH 0 lo0 4136
192.169.1.240/255.255.255.255 192.169.1.240 UH 0 lan1 4136
172.30.8.197/255.255.255.255 172.30.8.197 UH 0 lan0 4136
172.30.8.0/255.255.255.0 172.30.8.197 U 2 lan0 1500
127.0.0.0/255.0.0.0 127.0.0.1 U 0 lo0 0
default/0.0.0.0 172.30.8.1 UG 0 lan0 0

[Laurent Menase]

camracd0:>netstat -rnv
Routing tables
Dest/Netmask Gateway Flags Refs Interface Pmtu
127.0.0.1/255.255.255.255 127.0.0.1 UH 0 lo0 4136
192.169.1.240/255.255.255.255 192.169.1.240 UH 0 lan1 4136
172.30.8.197/255.255.255.255 172.30.8.197 UH 0 lan0 4136
172.30.8.0/255.255.255.0 172.30.8.197 U 2 lan0 1500
127.0.0.0/255.0.0.0 127.0.0.1 U 0 lo0 0
default/0.0.0.0 172.30.8.1 UG 0 lan0 0

It is strange that you don't have any
192.169.1.0/255.255.255.0 192.168.1.240 U 2 lan1 1500
it is like if you configured it as a point to point.
can you send also a
netstat -ni and ifconfig lan1

you should make a
route add net 192.169.1.0 192.168.1.240

then any message sent to 192.169.1.* will be send through lan1
all other traffic should gor through lan0

[Idham]

Thanks for the reply.

BTW..below is the sample traceroute ..
Which still use lan 1 instead of lan0 raouting

---/etc/hosts------------

172.30.8.197 camracd0
192.169.1.240 camracd0 #nbu
127.0.0.1 localhost loopback

---traceroute----
camracd0:>traceroute 172.30.8.197
traceroute: Warning: Multiple interfaces found; using 192.169.1.240 @ lan1
traceroute to 172.30.8.197 (172.30.8.197), 30 hops max, 40 byte packets
1 camracd0 (172.30.8.197) 0.082 ms 0.046 ms 0.014 ms

---netstat -in

camracd0:>ifconfig lan1
lan1: flags=1842
inet 192.169.1.240 netmask ffffff00 broadcast 192.169.1.255
camracd0:>netstat -in
Name Mtu Network Address Ipkts Ierrs Opkts Oerrs Coll
lan1* 1500 192.169.1.0 192.169.1.240 141761 6 319 0 0
lan0 1500 172.30.8.0 172.30.8.197 783607563 0 94770870 0 0
lo0 4136 127.0.0.0 127.0.0.1 29784553 0 29784553 0 0

[V. Nyga]

Hi,

check with 'lanscan' if lan0 is 'UP' or 'DOWN'.
Can you give us the output?

Volkmar

[Idham]

camracd0:>lanscan
Hardware Station Crd Hdw Net-Interface NM MAC HP-DLPI DLPI
Path Address In# State NamePPA ID Type Support Mjr#
0/1/2/0 0x001438EB9BD4 0 UP lan0 snap0 1 ETHER Yes 119
0/1/2/1 0x001438EB9BD5 1 UP lan1 snap1 2 ETHER Yes 119
0/3/1/0 0x001321BD3E72 2 UP lan2 snap2 3 ETHER Yes 119
0/4/1/0 0x001321BD3EC9 3 UP lan3 snap3 4 ETHER Yes 119

[V. Nyga]

Hi again,

somewhere you have a typo error:
192.169.1.240/255.255.255.255 192.169.1.240 UH 0 lan1 4136

You want
192.169.1.40/255.255.255.255 192.169.1.240 UH 0 lan1 4136

240 - 40

Where have you defined the routing?
'/etc/rc.config.d/netconf' ?

V.

[Idham]

Sorry ..
I deleted the 192.169.1.40 routing before I capture the netstat -rn. I need to delete it because dont want the production traffic use the 192.169.1.40 routing(Which is the main problem)


#route add host 192.169.1.40 192.169.1.24

camracd0:>netstat -rn
Routing tables
Destination Gateway Flags Refs Interface Pmtu
127.0.0.1 127.0.0.1 UH 0 lo0 4136
192.169.1.240 192.169.1.240 UH 0 lan1 4136
172.30.8.197 172.30.8.197 UH 0 lan0 4136
192.169.1.40 192.169.1.240 UH 0 lan1 0
172.30.8.0 172.30.8.197 U 2 lan0 1500
127.0.0.0 127.0.0.1 U 0 lo0 0
default 172.30.8.1 UG 0 lan0 0

Please help.

[V. Nyga]

also:
netstat -in
Name Mtu Network Address Ipkts Ierrs Opkts Oerrs Coll
lan1* 1500 192.169.1.0 192.169.1.240 141761 6 319 0 0

you have had errors at lan1

Can you reset the error counter and try it again? (landiag?)

V.

[Idham]

Hi V. Nyga

Can you tell me how to reset the counter. ?
Or maybe the full command.

FYI, we dont face similar issue on UNIX server running other than HPUX OS.

[V. Nyga]

Hi again,

my example if for hp-ux 11.11:

'landiag'
-> Test Selection mode.

lan = LAN Interface Administration
menu = Display this menu
quit = Terminate the Administration
terse = Do not display command menu
verbose = Display command menu

'lan'
ppa = PPA Number of the LAN Interface

'ppa 1'

'dis' -> Display LAN Interface status and statistics registers:
second page:

Ethernet-like Statistics Group

Index = 1
Alignment Errors = 0
FCS Errors = 0
Single Collision Frames = 0
Multiple Collision Frames = 0
Deferred Transmissions = 0
Late Collisions = 0
Excessive Collisions = 0
Internal MAC Transmit Errors = 0
Carrier Sense Errors = 0
Frames Too Long = 0
Internal MAC Receive Errors = 0

'clear' -> Clear statistics registers

-> all counters of 'dis' are zero.

Test lan1 again - new errors?
Any traffic at all?

What's about disabling production lan0 ?
Traffic at lan1 ?

Maybe the switch is not configured properly?
Half-Duplex/Full-Duplex?

V.

[Laurent Menase]

in fact everything is in:
camracd0:>traceroute 172.30.8.197
traceroute: Warning: Multiple interfaces found; using 192.169.1.240 @ lan1


When you have more than one interface on a system, traceroute is chosing one to make its probe, whatever be the routing table!

So if you want to check traceroute is not the right tool to use except if you precise the -i option.

try ping -o 172.30.8.197 -n 3

[V. Nyga]

Hi again,

sorry, I think my last answers go to the wrong direction lan1 is used the most time and you don't want to, right?

Another try:
Where's the backup server '192.169.1.40' in your host?
Also you check traceroute to yourself (172.30.8.197).

What's about another (external) ip?

V.

[Laurent Menase]

traceroute chose one interface but don't look at the routing table to chose it, so
don't use traceroute to do that.


in your case traceroute chose 192.169.1.240
to send its traffic.
but if you telnet, ping or watever you use it will chose according to the routing table.


try ping -o you will see that.

or make a network trace with tcpdump for instance, you will see that the interface used is lan1 only for 192.169.1.*

Don"t use traceroute to check that on a multi homed system!

[Idham]

Yes you are correct!!
"Don"t use traceroute to check that on a multi homed system!"

But it in a real time on our production server...the traffic went hywire....
whenever we enable the netbackup routing..

We already report this case to hp last week...
and they gave this kind of solution

1. ndd -set /dev/ip ip_strong_es_model 0
which require "PHNE_37670 " patch....
anyone have tried install this patches..
any problem so far?
FYI...our production server is running Oracle 10g RAC.

[rick jones]

Rather than zero counters and (possibly) confuse SNMP one can instead use beforeafter from ftp://ftp.cup.hp.com/dist/networking/tools/ to "subtract" one lanadmin -b mibstats or netstat -s output from another:

netstat -s -p tcp > before

netstat -s -p tcp > after
beforeafter before after > delta
less delta

etc etc

I'm guessing you have remote clients in other subnets and that is why return traffic to them would be following the default route even though you have your production and backup interfaces configured into separate IP subnets.

You will need to add an additional "default" route pointing at an IP in your backup subnet range and then indeed set ip_strong_es_model to either 1 or 2. I would suggest experiementing with this at a time when the DB is not in active production use, just to be careful.

You also need to be sure that if your backup server software _initiates_ connections (rather than passively accepts them) that it will bind to your backup network IP directly otherwise route selection may cause it to use the production IP and flow through the production network.