Hello,
I’m getting complaints from my family “the internet does not work”…
Short:
Today I was able to detect name resolution on a laptop connected on ETH to my Turris 1.0, while ping to internet address works:
root@:~# nslookup us-east-2.console.aws.amazon.com
;; connection timed out; no servers could be reached
dig:
root:~# dig us-east-2.console.aws.amazon.com
; <<>> DiG 9.11.19 <<>> us-east-2.console.aws.amazon.com
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: SERVFAIL, id: 6768
;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 8192
;; QUESTION SECTION:
;us-east-2.console.aws.amazon.com. IN A
;; Query time: 2 msec
;; SERVER: 127.0.0.1#53(127.0.0.1)
;; WHEN: Mon Jan 31 20:03:39 CET 2022
;; MSG SIZE rcvd: 61
ping:
root:~# ping 217.31.205.50
PING 217.31.205.50 (217.31.205.50): 56 data bytes
64 bytes from 217.31.205.50: seq=0 ttl=56 time=39.398 ms
64 bytes from 217.31.205.50: seq=1 ttl=56 time=21.014 ms
64 bytes from 217.31.205.50: seq=2 ttl=56 time=19.195 ms
^C
--- 217.31.205.50 ping statistics ---
3 packets transmitted, 3 packets received, 0% packet loss
round-trip min/avg/max = 19.195/26.535/39.398 ms
repeat the dig:
root:~# dig us-east-2.console.aws.amazon.com
; <<>> DiG 9.11.19 <<>> us-east-2.console.aws.amazon.com
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: SERVFAIL, id: 61140
;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 8192
;; QUESTION SECTION:
;us-east-2.console.aws.amazon.com. IN A
;; Query time: 2 msec
;; SERVER: 127.0.0.1#53(127.0.0.1)
;; WHEN: Mon Jan 31 20:05:32 CET 2022
;; MSG SIZE rcvd: 61
and old school nslookup:
;; connection timed out; no servers could be reached
So I logged to forris and disabled DNSSEC and name resolution work:
root:~# nslookup us-east-2.console.aws.amazon.com
Server: 127.0.0.1
Address: 127.0.0.1#53
Name: us-east-2.console.aws.amazon.com
us-east-2.console.aws.amazon.com canonical name = gr.console-geo.us-east-2.amazonaws.com
Name: gr.console-geo.us-east-2.amazonaws.com
gr.console-geo.us-east-2.amazonaws.com canonical name = af2049b9c08c62706.awsglobalaccelerator.com
Name: af2049b9c08c62706.awsglobalaccelerator.com
Address 1: 76.223.79.155
Address 2: 13.248.199.77
us-east-2.console.aws.amazon.com canonical name = gr.console-geo.us-east-2.amazonaws.com
gr.console-geo.us-east-2.amazonaws.com canonical name = af2049b9c08c62706.awsglobalaccelerator.com
root@molekula:~# dig us-east-2.console.aws.amazon.com
; <<>> DiG 9.11.19 <<>> us-east-2.console.aws.amazon.com
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 1786
;; flags: qr rd ra; QUERY: 1, ANSWER: 4, AUTHORITY: 0, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 8192
;; QUESTION SECTION:
;us-east-2.console.aws.amazon.com. IN A
;; ANSWER SECTION:
us-east-2.console.aws.amazon.com. 59 IN CNAME gr.console-geo.us-east-2.amazonaws.com.
gr.console-geo.us-east-2.amazonaws.com. 60 IN CNAME af2049b9c08c62706.awsglobalaccelerator.com.
af2049b9c08c62706.awsglobalaccelerator.com. 300 IN A 76.223.79.155
af2049b9c08c62706.awsglobalaccelerator.com. 300 IN A 13.248.199.77
;; Query time: 752 msec
;; SERVER: 127.0.0.1#53(127.0.0.1)
;; WHEN: Mon Jan 31 20:06:12 CET 2022
;; MSG SIZE rcvd: 195
root:~# cat /etc/turris-version
3.11.23
Sometime later (I’ve found in the forums options like +trace
so I’ve enabled DNSSEC back and surprisingly, it worked.
One more thing I did, while the name resolution with DNSSEC does not work on my Turris 1.0, I’ve tried to test the name resolution using Omnia (TOS5, DNSSEC enabled) and it properly worked.
All posts to DNSSEC I’ve found were about bad answers for a domain (BNP Paribas…), but this looks different to me as one Turris was able to resolve the name with DNSSEC, while the other was not.
Honestly, to me, it looks the same issue happens on Omnia with TOS5 (and previously TOS3) as user symptoms are the same, but this time I was able to connect to Turris and do some tests.
The question is, how to configure some diagnostic to capture log files that can be analyzed, I hardly believe that two independent internet providers will suffer from the same (or that too similar) issue(s).
The intermittent issue is unlikely to be captured by and analyzed hand…
From a configuration point of view, regarding DNS I only used Forris interface to (initially enable DNSSEC + forwarding to CZ NIC servers) disable forwarding and for the test case also to disable DNSSEC (and finally re-enable it back), so no UCI commands nor configuration file manual edits.
Long story:
What I was able to observe, switching WiFi on android mobile off and on resolves that (bloody android ?.. well, it looks no). Smart TV sometimes freezes (usually when ads break the movie…).
Than bad part is, if that happens while using mobile banking…
Having a laptop connected to the same wifi, ping to a known address from the internet works and rarely we are above 10Mbps while using Turris 1.0 with provider claiming (and according to netmeter it is true) 100Mbps downlink (provide use a 5G modem) and a second provider connected to Turris Omnia (recently upgraded to TOS5) with symmetric connectivity 20Mbps (according to netmetr, it works as well).
The only suspect in my mind (meaning I’ve ignored all more probable options…) was DNS resolving.
None of my providers supports DNSSEC, so I’ve been using forwarding to CZ.NIC validating resolvers, but since one day I received an email from my router that it seem not to work properly and suggested to switch off forwarding, I’m not using forwarding now.
Thanks for any help, Ales