Skip to contentSkip to conversation

Route works, but ping time is 1000 ms

LARTC Philip Mak 2 participants 4 messages Feb 14, 2002 – Feb 15, 2002

lore
Share & tools
lore

Philip Mak

author
#0 lore raw

I have a server that has two ethernet cards, eth0 (64.94.46.199) and eth1 (66.28.98.52).

eth0's ISP has expensive, high quality bandwidth.
eth1's ISP has cheap, low quality bandwidth.

I have the route set so that by default, all packets will go out through eth1:

default via 66.28.98.34 dev eth1

However, I'd like to make it so that when I ssh to 64.94.46.199 (eth0, the expensive high quality ISP), the return traffic will go out through eth0. So, I typed this:

$ echo "200 pnap" >> /etc/iproute2/rt_tables
$ ip rule add from 64.94.46.199 lookup pnap
$ ip route add default via 64.94.47.253 dev eth0 table pnap

This correctly makes it route outgoing packets with source IP of 64.94.46.199 through eth0.

However, the ping time suddenly goes up a lot on the last hop (you can traceroute 64.94.46.199 and see for yourself).

Any idea what I did wrong? I have a hunch that my ISP's router doesn't like some settings in the packets I sent to it and is giving me low priority as a result, but I wouldn't know how to tell if that was the case... (I e-mailed my ISP for help on Tuesday; still waiting.)
_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/lartc/

#1 lore raw
Quoted 2 lines · Philip Mak

I have a server that has two ethernet cards, eth0 (64.94.46.199) and eth1 (66.28.98.52).

Can you show what happens? So something like this:

# ping some.ip.address.0
ms/ms/ms/ms
# echo "200 pnap" >> /etc/iproute2/rt_tables
# ip rule add from 64.94.46.199 lookup pnap
# ip route add default via 64.94.47.253 dev eth0 table pnap
# ping some.ip.address.0
ms/ms/ms/ms

Preferably with tcpdumps illustrating where packets go out.

Regards,

bert

Signature · 7 lines
http://www.PowerDNS.com          Versatile DNS Software & Services
http://www.tk                              the dot in .tk
Netherlabs BV / Rent-a-Nerd.nl           - Nerd Available -
Linux Advanced Routing & Traffic Control: http://ds9a.nl/lartc
_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/lartc/
author
#2 lore raw
In reply to Philip Mak
Quoted 4 lines · bert hubert
Quoted 2 lines

I have a server that has two ethernet cards, eth0 (64.94.46.199) and eth1 (66.28.98.52).

Can you show what happens? So something like this:

I found out that even before I make the IP routing table changes that I mentioned in my previous message, there is already this problem.

Check this out:

[root@lina rc.d]# ping -r -b 64.94.47.255
WARNING: pinging broadcast address
PING 64.94.47.255 (64.94.47.255) from 64.94.46.199 : 56(84) bytes of data.
Warning: time of day goes back, taking countermeasures. 64 bytes from 64.94.46.199: icmp_seq=0 ttl%5 time=1.081 msec 64 bytes from 64.94.46.199: icmp_seq=1 ttl%5 time9 usec 64 bytes from 64.94.46.35: icmp_seq=0 ttl%5 time=1.012 sec (DUP!) 64 bytes from 64.94.46.206: icmp_seq=0 ttl%5 time=1.013 sec (DUP!) 64 bytes from 64.94.46.197: icmp_seq=0 ttl%5 time=1.013 sec (DUP!) 64 bytes from 64.94.46.202: icmp_seq=0 ttl%5 time=1.013 sec (DUP!) 64 bytes from 64.94.46.200: icmp_seq=0 ttl%5 time=1.014 sec (DUP!)

The above is a ping broadcast. I used the -r flag to make ping send directly to hosts on the local network, bypassing the routing tables (so that I know my routing tables are not an issue). But yet, the ping times are 1 second which is outrageous for hosts on a local network.

Any ideas what I should try next? Would this be a problem with the ethernet card, network cable, switch, or any of the above?
_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/lartc/

author
#3 lore raw
In reply to Philip Mak

I'm not sure what I did, but everything seems to be working correctly now.

Well, I still have one problem: If the machine reboots, I lose all my settings. Can anyone point me to a good reference of how to make these settings take effect at boot time? I assume there's a better way other than putting the commands into /etc/rc.d/rc.local.

Script started on Fri Feb 15 13:34:13 2002 [root@lina route]# tail -1 /etc/iproute2/rt_tables 200 pnap
[root@lina route]# ip route show 66.28.98.32/27 via 66.28.98.34 dev eth1 66.28.98.32/27 dev eth1 scope link 64.94.46.0/23 dev eth0 scope link 64.94.46.0/23 via 64.94.47.253 dev eth0 127.0.0.0/8 dev lo scope link
default via 66.28.98.34 dev eth1 [root@lina route]# ip rule show

0:      from all lookup local 
32765:  from 64.94.46.199 lookup pnap 
32766:  from all lookup main 
32767:  from all lookup 253 

[root@lina route]# ip route show table pnap
default via 64.94.47.253 dev eth0
[root@lina route]#
Script done on Fri Feb 15 13:34:40 2002
_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/lartc/