0

I am trying to get several ip addresses to be attached to a virtual server running ubuntu 10.04 on vmware esxi.

using esxi i have assigned it 10 virtual nic's which are all connected to the same virtual switch.

each virtual nic has its own ip address which no other device on the network is using.

The problem is that from within the office, i.e. before the router everything works just fine no problems, all services can be used just fine. from outside the office, i.e. public internet only one ip on the machine provides any services. that is the ip on eth0. all the other ip's time out.

all ip's are public ip's in the same range. There is no firewall issue because the firewall configuration has not been changed, and all ip's were publicly available on the old hardware that has been powered off.

traceroute outputs from another server elsewhere are as follows:

root@jaguar:~# tcptraceroute x.x.x.252
traceroute to x.x.x.252 (x.x.x.252), 30 hops max, 60 byte packets
 1  * * *
 2  * * *
 3  Xe0-1-1-0-grtontl1.red.telefonica-wholesale.net (213.140.51.181)  21.845 ms  22.682 ms  23.545 ms
 4  GE6-1-0-0-grtlontl3.red.telefonica-wholesale.net (84.16.12.14)  24.144 ms  24.834 ms GE7-0-0-0-grtlontl3.red.telefonica-wholesale.net (84.16.12.94)  25.374 ms
 5  CW-2-1-0-0-grtlontl3.red.telefonica-wholesale.net (213.140.52.62)  67.339 ms  67.334 ms  67.327 ms
 6  xe-4-2-0-xcr1.lnd.cw.net (195.2.25.58)  86.102 ms  64.355 ms  64.136 ms
 7  ge-5-0-0-dcr1.lnd.cw.net (195.2.25.2)  18.486 ms ge-1-0-0-dcr1.lnd.cw.net (195.2.25.122)  18.664 ms  18.613 ms
 8  so-0-0-0-zcr1.lnt.cw.net (166.63.163.222)  18.830 ms so-4-0-0-zcr1.lnt.cw.net (166.63.161.146)  18.437 ms so-0-0-0-zcr1.lnt.cw.net (166.63.163.222)  18.372 ms
 9  so-1-3-0-ycr1-lo7.cw.net (166.63.222.30)  22.153 ms  22.413 ms  22.411 ms
10  ge-0-0-0-kar1.lo7.cw.net (195.2.12.178)  22.886 ms  22.404 ms  22.356 ms
11  hackney-gw.uk.insnet.net (213.38.241.146)  28.065 ms  28.818 ms  29.410 ms
12  * * *
13  * * *
14  * * *
15  * * *
16  * * *
17  * * *
18  * * *
19  * * *
20  * * *
21  * * *
22  * * *
23  * * *
24  * * *
25  * * *
26  * * *
27  * * *
28  * * *
29  * * *
30  * * *
root@jaguar:~# tcptraceroute x.x.x.244
traceroute to x.x.x.244 (x.x.x.244), 30 hops max, 60 byte packets
 1  * * *
 2  * * *
 3  * * *
 4  GE7-0-0-0-grtlontl3.red.telefonica-wholesale.net (84.16.12.94)  23.450 ms  24.431 ms GE6-1-0-0-grtlontl3.red.telefonica-wholesale.net (84.16.12.14)  24.956 ms
 5  CW-2-1-0-0-grtlontl3.red.telefonica-wholesale.net (213.140.52.62)  25.626 ms  26.377 ms  26.975 ms
 6  xe-4-0-0-xcr1.lnd.cw.net (195.2.25.30)  27.823 ms xe-4-2-0-xcr1.lnd.cw.net (195.2.25.58)  18.397 ms  18.448 ms
 7  ge-5-0-0-dcr1.lnd.cw.net (195.2.25.2)  18.660 ms ge-1-0-0-dcr1.lnd.cw.net (195.2.25.122)  22.714 ms  21.367 ms
 8  so-4-0-0-zcr1.lnt.cw.net (166.63.161.146)  18.789 ms so-0-0-0-zcr1.lnt.cw.net (166.63.163.222)  18.915 ms so-4-0-0-zcr1.lnt.cw.net (166.63.161.146)  18.532 ms
 9  so-1-3-0-ycr1-lo7.cw.net (166.63.222.30)  22.877 ms  22.694 ms  22.405 ms
10  ge-12-0-0-kar1.lo7.cw.net (195.2.12.182)  22.404 ms  22.430 ms  22.600 ms
11  hackney-gw.uk.insnet.net (213.38.241.146)  27.783 ms  27.810 ms  28.225 ms
12  rsvd-company-name-244.x.x.x.in-addr.arpa (x.x.x.244)  28.292 ms  28.295 ms  27.762 ms

the address ending in 252 is one that does not work the address ending in 244 is one that does work

/etc/network/interfaces looks like:

auto lo
iface lo inet loopback

auto eth0
iface eth0 inet static
    address x.x.x.244
    netmask 255.255.255.0
    broadcast x.x.x.255
    gateway x.x.x.x
    dns-nameservers x.x.x.x

auto eth1
iface eth1 inet static
    address x.x.x.252
    netmask 255.255.255.0
    gateway x.x.x.x
    broadcast x.x.x.255
...

etc etc

4 Answers 4

1

Finally found the solution.

It was not anything to do with the configuration on Ubuntu or VMWare but was in fact the router itself. It's arp table seems to have an extremely long lifetime. Power cycling the router solved the problem, and has solved the problem every time it has occured since then.

Just posting this so that anyone else who might find this situation may find some help.

0

Are all the IP addresses on the same subnet?

If so:

  • you might want to use virtual interfaces in ubuntu rather than real ones (ie: eth0, eth 0:1, eth0:2, etc) to reduce the complexity of your virtual switch configuration
  • you only need one gateway defined

If not, what's probably hosing you is the fact that you are defining 10 gateways which may or may not be the same; and which may or may not permit "source routing" (ie accepting outbound packets from sources which don't match the incoming routing rules). In that case, you want to google "source routing".

1
  • i had already tried it without each having a gateway and also with virtual interfaces. it was however working with its current setup until i rebooted. i had previously rebooted, so it wasn't just a change in settings that had done it.
    – m3z
    Aug 13, 2010 at 7:56
0

Check the switch configuration inside of VMware. Also try using a ping -I and select the interface that is not working to see if that resolves the problem. (It may be that it just does not have an arp entry.

1
  • thanks, next time i come across something like this i'll be sure to try that
    – m3z
    Aug 13, 2010 at 7:57
0

Not sure what the problem was exactly, but after i changed all the interfaces into virtual interfaces on eth0, shutdown, removed the extra nic's from the esxi virtual machine settings and booted up again it was all working.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .