0

We have 2 physical machines, Host1 is a CentOS 6.4 kvm host and hosts ~7 kvm VMs all running Ubuntu 12.04 - all of this runs perfectly.

Recently we've added a 2nd host system, host2, again a CentOS 6.4 kvm host with a view to running another couple of VMs and providing some failover against host1 should it be required.

Both physical machines reside in the same cabinet in our DC, and are on the same subnet - let's say host1: 1.1.1.64 and host2: 1.1.1.81. Both have their gateway set to the DC gateway of 1.1.1.254 with no hardware firewall in between.

On each machine, I have 4 NICs that are bonded together to form a single interface, which is then bridged to allow the VMs to access the network. All of the VMs are online, and all of them can successfully ssh into the hosts without any delay.

Both systems can access the internet fine, and I can ssh into both systems from home without any issues. However, there is a real delay when attempting to ssh from host1 to host2 (or vice versa) and this obviously means that any action required on host2, that is controlled by host1 either takes forever or results in failure due to timeout.

In the interest of keeping this post short, I've put my ifcfg files into a pastie: http://pastie.org/8081648

I've tried both adding a firewall rule in each machine for the other, and also disabling the firewall entirely, so that can't be the issue.

I've tried troubleshooting this myself but can't seem to get to the bottom of it. Any help or advice would be appreciated.

Thanks in advance.

  • It's not a good idea to use other people's public IP addresses for your examples. See [RFC 5737](http://tools.ietf.org/html/rfc5737). – Michael Hampton Jun 26 '13 at 13:51

1 Answers1

0

The problem was related to the DNS settings, the IP * .* .* .240 was incorrect, so whenever an ssh connection was initiated and the reverse lookup requested from this DNS address the response timed out and caused the delay.

So if you have ssh login delays - check domain name resolution from the server.