0

If I check how many connections serverA (192.168.1.1) has open to serverB (192.168.2.1), I get the following response:

[username@serverA ~] $ netstat -n | grep 192.168.2.1
tcp        0      0 192.168.1.1:51846     192.168.2.1:10001     ESTABLISHED
tcp        0      0 192.168.1.1:50872     192.168.2.1:10001     ESTABLISHED
tcp        0      0 192.168.1.1:51824     192.168.2.1:10001     ESTABLISHED
tcp        0      0 192.168.1.1:51848     192.168.2.1:10001     ESTABLISHED
[username@serverA ~] $ netstat -n | grep 10.79.165.145 | wc -l
4

However, if I do the opposite and check on serverB how many connections it has open to serverA, I get this:

[username@serverB ~] $ netstat -n | grep 192.168.1.1
tcp        0      0 192.168.2.1:10001     192.168.1.1:51846     ESTABLISHED
tcp        0      0 192.168.2.1:10001     192.168.1.1:55122     ESTABLISHED
tcp        0      0 192.168.2.1:10001     192.168.1.1:59930     ESTABLISHED
tcp        0      0 192.168.2.1:10001     192.168.1.1:50352     ESTABLISHED
tcp        0      0 192.168.2.1:10001     192.168.1.1:44142     ESTABLISHED
tcp        0      0 192.168.2.1:10001     192.168.1.1:57698     ESTABLISHED
tcp        0      0 192.168.2.1:10001     192.168.1.1:38268     ESTABLISHED
tcp        0      0 192.168.2.1:10001     192.168.1.1:41822     ESTABLISHED
                    ... many more connections ... 
tcp        0      0 192.168.2.1:10001     192.168.1.1:43840     ESTABLISHED
tcp        0      0 192.168.2.1:10001     192.168.1.1:50870     ESTABLISHED
tcp        0      0 192.168.2.1:10001     192.168.1.1:34100     ESTABLISHED
tcp        0      0 192.168.2.1:10001     192.168.1.1:34620     ESTABLISHED
tcp        0      0 192.168.2.1:10001     192.168.1.1:41126     ESTABLISHED
tcp        0      0 192.168.2.1:10001     192.168.1.1:49298     ESTABLISHED
tcp        0      0 192.168.2.1:10001     192.168.1.1:50004     ESTABLISHED
tcp        0      0 192.168.2.1:10001     192.168.1.1:51408     ESTABLISHED
[username@serverB ~] $ netstat -n | grep 192.168.1.1 | wc -l
104

I was not expecting there to be a mismatch in the number of the connections between the 2 servers, essentially serverB thinks there are a lot more connections open to serverA and than that serverA does to serverB.

These servers are on different VLANs and the connections do go through a firewall. Both serverA and serverB are RHEL v7 VMs running on ESXI. They don't run any containers or anything that would do NAT.

What could be responsible for the mismatch in open connection numbers

2
  • 1
    Is serverA a router doing NAT? Running containers or VMs also counts as a router doing NAT.
    – A.B
    Commented May 15, 2020 at 14:23
  • @A.B Nope, nothing special going on in terms of networking, just plain old VMs with static IPs, and we don't use docker or any containers on them. Updated post with this info.
    – Ben Dyson
    Commented May 15, 2020 at 17:07

1 Answer 1

0

This was solved by a patch release for the application server in use.

You must log in to answer this question.

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