asked on

I cannot ping an esx host server from a different server

I cannot ping an esx host server from a different server (call it server1)

1. I can ping BOTH the esx host and server1 from my laptop.
2. This is relativley new.
3. I turned the firewall off on server1 with no effect.
4. There have been no changes to the esx server.
5. I tried pinging server1 from the esx host via ssh and was unable to get a response.
6. There does not seem to be any networking issues. I can ping the gateway's and other end points.
7. It seems like server1 is not responding to ICMP requests, but with the firewall off, I am at a loss.

Gerald Negrota

On server1 add a firewall rule that permit ping. After that :
Check list:
- change cables with a known good ones ;
- verify that you can ping default gw of both ;
- verify that you have the same VLAN or both VLAN are permitted rules ;
- routing table on hosts and router/switch are ok;
- ethernet interface are compatible with switch (try vary speed fixed);
Next actions will be in debug mode... Let me know.

Glen Krinsky

ASKER

I added the firewall rule, even though the firewall is turned of...just in case. no joy. I can ping everything else, both ways, from both sides, EXCEPT to each other. I am seeing nothing in any event log. We had this issue before and it was firewall related (Same servers) and we were able to fix it.

Gerald Negrota

Let the firewall on with proper rules. Other points in check list ?!
If both are on different networks, for me is a routing/switching problem. Verify routing table on both.

Glen Krinsky

ASKER

I have tried with the firewall on as well. And yes, the routing tables are correct.

Rich Weissler

Assumptions (please comment/correct incorrect assumptions):
- We're working entirely in IPv4, there isn't any IPv6 configured.
- No NATing is taking place.
- Subnet masks are also correct on server1, the esx host, AND all router interfaces.
- There is no default gateway redundancy protocols in operation.

Are you running NSX in your VMWare environment?
Are server1 and the ESX host on the same subnet or different subnets?
Does your laptop share a subnet with either of other two machines?
How many switches (virtual or physical) are on each of subnets?

Can you provide a sketch of the network configuration with the laptop, Server1, ESX host, routers? (And ideally any VLANs and VXLANs.)

Glen Krinsky

ASKER

Your assumptions are correct. The esx host is on a different network (10.17.x.x) than server1 and my laptop(10.3.x.x). There is a single switch on each side. This was working up until a few weeks ago. There have been zero changes to the network which is why I do not believe it to be a network issue (I can also ping other devices on the same network as esx host from server 1).

Rich Weissler

The problem is because there was a behavior change, there was has been some sort of change somewhere. It may not be an intentional change. It might be a change caused because of an automated fault recovery, or an unrelated network cable change which caused changes (like spanning tree... but it doesn't sound like STP is the culprit in this case. It might be a software patch installed somewhere.
It's possible it isn't network related. It might not be protocol stack related. But I believe you're eliminated everything you believe it to be... so we're left looking at the things you don't believe it to be. :-) (otherwise it'd already be solved.)

Do I read correctly that server1 and your laptop are on subnet 10.3.x.x? Do both of those have subnet masks of 255.255.0.0? (If server1 accidently picked up a mask of 255.0.0.0... I could understand the problem... and 255.0.0.0 would be the default.)

ASKER CERTIFIED SOLUTION

Glen Krinsky

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

Gerald Negrota

Last resort, restart (only) management interface from esx console.

Gerald Negrota

DNS has nothing to do with ping <host-ip> ...
Anyway, glad to hear that is fixed.