03-08-2015 10:44 AM - edited 03-07-2019 10:59 PM
Hi
I have a server, that monitors the virtual hosts by send pings to them. When a virtual host is down, the server will send out an alert. This server and hosts are currently on VLAN 1.
After upgrading the core network, I noticed the alerts not working. When shutting down a virtual host, I received no alerts. After some investigation I shutdown the virtual host and tried pinging it from the server and obviously expected the pings to fail but it was coming back with "request timed out for the first few pings and then "destination host unreachable". When I bring the virtual host back up it pings as normal.
So I believe the reason I am not receiving the alerts is because the ping is not timing out but instead coming back with destination host unreachable.
So I moved the server and virtual hosts to a different vlan and subnet and now everything works and pings time out when i shut down any virtual host.
So I am trying to understand why on vlan 1 the message returned destination host unreachable and on a different vlan it came back with request timed out.
I have found posts that tell me destination host unreachable means the sender has no route to the destination, but when i bring up the destination it all of a sudden can ping and gets that route. A bit confusing.
Any help will be appreciated
Thanks
03-08-2015 02:45 PM
I have found posts that tell me destination host unreachable means the sender has no route to the destination, but when i bring up the destination it all of a sudden can ping and gets that route. A bit confusing.
It does mean that but perhaps not in the way you think in terms of "route".
Do you remember whether the destination host unreachable message was reported by the switch or not ?
If everything is in the same vlan then there is no actual routing to be done ie. it is all L2 and not L3.
So what it is telling you is that the server could not resolve the IP address to a mac address. If it has no mac address for the host it cannot put any packets on the wire so it reports destination host unreachable.
With a request timed out packets can be put on the wire and sent but there is no response obviously because the host is down. So it's almost as if the server has a mac address for the first few pings and then removes it from it's arp cache and then cannot send any more packets.
When you bring the host back up the server can then get the mac address through arp and your pings start working again.
What this has to do with the switch I'm not sure because it should not be involved. It's not about the switches mac address table because even if the mac address wasn't there that is at L2 so the switch should not generate any message. It should only generate the message if it was being asked to L3 switch the packet and it isn't.
It's probably too late to do any troubleshooting as you have moved it all into a different vlan and it is working.
I can't think of any obvious reason as to why it behaved like this but just thought I would explain some of the symptoms you were seeing.
If anything does occur I'll post back.
Jon
03-08-2015 04:10 PM
Thanks Jon for the explanation, nice and clear. The response was coming back from the server itself so it was dropping the mac address after a few pings failed.
Although it worked by moving to a different vlan I was just curious as to why it was behaving like this on the default vlan.
Thanks
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide