frequent issue with routing from EC2 instance

Question

I have a problem with routing from my EC2 in eu-central-1 to a host in the cloud in Slovakia. EC2 instance is connected to public subnet and I have 1 elastic IP associated with this instance. Host in Slovak cloud is golan-api.gospace.tech. We send a lot of HTTP requests and I often get a "no route to host" error

If everything works fine traceroute looks like this
```
traceroute to golan-api.gospace.tech (185.111.90.73), 30 hops max, 60 byte packets
 1  ec2-54-93-0-77.eu-central-1.compute.amazonaws.com (54.93.0.77)  9.629 ms ec2-99-150-19-93.eu-central-1.compute.amazonaws.com (99.150.19.93)  8.367 ms ec2-99-150-19-77.eu-central-1.compute.amazonaws.com (99.150.19.77)  3.153 ms
 2  240.0.88.13 (240.0.88.13)  0.363 ms 240.0.88.14 (240.0.88.14)  0.234 ms 240.0.88.15 (240.0.88.15)  0.218 ms
 3  242.1.84.37 (242.1.84.37)  0.903 ms 242.1.85.165 (242.1.85.165)  0.762 ms 242.1.85.35 (242.1.85.35)  1.436 ms
 4  240.0.92.34 (240.0.92.34)  1.250 ms 240.0.96.35 (240.0.96.35)  0.822 ms 240.0.92.35 (240.0.92.35)  1.563 ms
 5  242.1.100.37 (242.1.100.37)  1.644 ms 242.1.93.161 (242.1.93.161)  1.072 ms 242.1.101.165 (242.1.101.165)  1.469 ms
 6  100.100.2.74 (100.100.2.74)  1.443 ms 100.100.4.68 (100.100.4.68)  1.062 ms *
 7  ffm-b16-link.ip.twelve99.net (62.115.58.114)  1.074 ms ffm-b18-link.ip.twelve99.net (62.115.51.84)  0.904 ms ffm-b11-link.ip.twelve99.net (62.115.48.54)  1.540 ms
 8  ffm-bb2-link.ip.twelve99.net (62.115.132.210)  1.756 ms ffm-bb2-link.ip.twelve99.net (62.115.124.118)  1.109 ms  1.199 ms
 9  win-bb1-link.ip.twelve99.net (62.115.137.203)  12.637 ms  12.614 ms win-bb2-link.ip.twelve99.net (62.115.138.23)  12.585 ms
10  brat-b2-link.ip.twelve99.net (62.115.121.223)  16.597 ms  16.798 ms  16.758 ms
11  lightstorm-ic-360602.ip.twelve99-cust.net (62.115.181.221)  16.745 ms  18.017 ms  16.849 ms
12  185.87.13.255 (185.87.13.255)  15.382 ms  15.369 ms *
13  73.90.111.185.in-addr.arpa.websupport.sk (185.111.90.73)  16.737 ms  16.663 ms *
```

Here are the outputs from ping, nslookup, traceroute and curl when there is a problem

Ping Output:
```
PING golan-api.gospace.tech (185.111.90.73) 56(84) bytes of data.
From 54.239.106.111 (54.239.106.111) icmp_seq=1 Time to live exceeded
From 54.239.106.111 (54.239.106.111) icmp_seq=2 Time to live exceeded
From 54.239.106.111 (54.239.106.111) icmp_seq=3 Time to live exceeded
From 54.239.106.111 (54.239.106.111) icmp_seq=4 Time to live exceeded

--- golan-api.gospace.tech ping statistics ---
4 packets transmitted, 0 received, +4 errors, 100% packet loss, time 303ms
```

nslookup Output:

```
Server:		127.0.0.53
Address:	127.0.0.53#53

Non-authoritative answer:
Name:	golan-api.gospace.tech
Address: 185.111.90.73
```

Traceroute Output:
```
traceroute to golan-api.gospace.tech (185.111.90.73), 30 hops max, 60 byte packets
 1  100.100.14.4 (100.100.14.4)  0.554 ms 100.100.14.112 (100.100.14.112)  1.448 ms 100.100.8.104 (100.100.8.104)  0.575 ms
 2  100.100.8.29 (100.100.8.29)  1.941 ms * *
 3  100.100.89.136 (100.100.89.136)  1.067 ms 100.100.68.206 (100.100.68.206)  1.195 ms 100.100.88.14 (100.100.88.14)  1.216 ms
 4  100.100.68.217 (100.100.68.217)  0.999 ms 100.100.92.89 (100.100.92.89)  1.254 ms  1.234 ms
 5  100.100.24.18 (100.100.24.18)  1.476 ms 100.100.24.32 (100.100.24.32)  13.171 ms 100.100.24.30 (100.100.24.30)  1.407 ms
 6  100.95.6.1 (100.95.6.1)  0.853 ms 100.95.6.49 (100.95.6.49)  10.856 ms 100.95.6.1 (100.95.6.1)  1.592 ms
 7  * * 150.222.194.32 (150.222.194.32)  1.701 ms
 8  54.239.106.83 (54.239.106.83)  1.085 ms 54.239.106.129 (54.239.106.129)  1.403 ms 54.239.106.173 (54.239.106.173)  1.020 ms
 9  * 100.106.108.135 (100.106.108.135)  0.926 ms 100.106.108.137 (100.106.108.137)  0.979 ms
10  100.100.14.3 (100.100.14.3)  1.246 ms 100.100.8.45 (100.100.8.45)  6.768 ms 100.100.14.53 (100.100.14.53)  7.185 ms
11  * 100.100.88.14 (100.100.88.14)  1.473 ms 100.100.89.8 (100.100.89.8)  1.343 ms
12  100.100.80.69 (100.100.80.69)  1.888 ms 100.100.81.67 (100.100.81.67)  1.519 ms 100.100.78.89 (100.100.78.89)  1.183 ms
13  100.100.24.50 (100.100.24.50)  2.054 ms 100.100.2.76 (100.100.2.76)  1.097 ms 100.100.4.74 (100.100.4.74)  1.488 ms
14  100.95.6.17 (100.95.6.17)  4.920 ms 100.95.6.49 (100.95.6.49)  6.842 ms 100.95.6.17 (100.95.6.17)  4.866 ms
15  150.222.194.36 (150.222.194.36)  1.698 ms ffm-bb1-link.ip.twelve99.net (62.115.124.116)  2.226 ms 150.222.194.32 (150.222.194.32)  1.799 ms
16  54.239.106.19 (54.239.106.19)  1.472 ms 54.239.106.155 (54.239.106.155)  1.489 ms win-bb1-link.ip.twelve99.net (62.115.137.203)  12.721 ms
17  100.92.18.83 (100.92.18.83)  4.015 ms * *
18  * * 100.100.14.23 (100.100.14.23)  1.605 ms
19  * 100.100.76.14 (100.100.76.14)  1.584 ms *
20  100.100.82.89 (100.100.82.89)  1.641 ms * *
21  100.100.24.34 (100.100.24.34)  1.420 ms * 100.100.2.90 (100.100.2.90)  1.347 ms
22  * * 100.95.6.1 (100.95.6.1)  1.519 ms
23  * 150.222.194.22 (150.222.194.22)  1.921 ms ffm-bb1-link.ip.twelve99.net (62.115.124.116)  2.068 ms
```

cURL Output:
```
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0
curl: (7) Failed to connect to golan-api.gospace.tech port 443 after 6 ms: No route to host
```
Could you advise me where the problem might be and how to solve it?

Ramakrishna Natarajan · Answer

Hi,

If the target system has no traffic filtering or limitations in place, it is advisable to contact AWS Enterprise Support for a detailed investigation of this specific issue.

Thanks,
Rama

Brettski · Answer

In the traceroute output when there is an issue, the initiating IP address is different to when your connection attempts work. Specifically, the source IP is 100.100.x.x which is a special (bogon) IP address range. Could it be that there's something different on the EC2 instance when things are not working?

EXPERT

Brettski

answered 2 years ago

skadlic · Answer

I tried deploying two new EC2 instances. One in the same region as the problematic one and the other in eu-west-3. The one in eu-central-1 shows the same problem. The EC2 instance in eu-west-3 did not show the problem.

I also tried the same trace routing towards the server in the Hetzner cloud from the problematic EC2. There the communication goes fine with no interruptions.

From Hetzner cloud to golan-api.gospace.tech I have also not experienced any problem so far.

frequent issue with routing from EC2 instance

Add your answer

Relevant content