[Solved] Random DNS Errors

Jocomol · 2023-08-15 17:00:57

Hello Y'all

I've recently installed Arch Linux + Larbs on my new T480s. Everything (except the kernel version, this one newest) is the same as my T14 (where I use LTS). But I've noticed that sometimes DNS isn't able to resolve. I've already checked if it happens on different networks and it does. I was unfortunately unable to find a issue that is the same as mine, often people experience package drops while it its only the DNS that seems to be cutting out sometimes.

CPU: Intel i5-8250U (8) @ 3.400GHz
Host: 20L7001VMZ ThinkPad T480s
- NetworkManager

Configured Nameservers:
- nameserver 8.8.8.8 (as a failover)
- nameserver 192.168.1.1 (DNS set by the DHCP of the network I am in)

I am glad for any help and happy to provide more info or logs.

Thank you very much!

Last edited by Jocomol (2023-08-25 15:17:30)

seth · 2023-08-15 18:25:53

Please post the output of

find /etc/systemd -type l -exec test -f {} \; -print | awk -F'/' '{ printf ("%-40s | %s\n", $(NF-0), $(NF-1)) }' | sort -f

Also

resolvectl status

When you've dns issues, try

nslookup <unresolvable_domain_here>
dig @8.8.8.8  <unresolvable_domain_here>

(both are part of the bind package) and post the outputs.

Jocomol · 2023-08-15 18:38:35

First and Second output
http://0x0.st/HLP1.md
I think the problem might be found in the output of the second command.

http://0x0.st/HLP_.md

I ran dig and nslookup on a loop, nslookup was able to resolve in the window when dig wasn't

seth · 2023-08-15 18:44:04

You're timing out to 8.8.8.8, that's not just a DNS issue.
nslookup likely gets it out of a local cache (depending on what the primary DNS is, maybe your gateway runs dnsmasq)
You're simply not using resolved (what's ok) and the obvious contender is the VPN. Stop/disable mullvad and see whether you can reproduce the issue w/o.

Jocomol · 2023-08-15 18:47:34

The VPN wasn't connected at the time of the issue and the issue occured even before I installed mullvad on this machine. In any case I've stopped the daemon and will try to see if I can reproduce the issue.

Jocomol · 2023-08-15 18:53:12

Yes, the issue still occures

seth · 2023-08-15 19:04:00

ip a; ip r

and we'll probably need a system journal after this has happened at least once,

sudo journalctl -b | curl -F 'file=@-' 0x0.st

for the current boot.

;; communications error to 8.8.8.8#53: timed out

isn't a misconfigured DNS.

Jocomol · 2023-08-15 19:06:37

ip a; ip r
http://0x0.st/HLP6.txt
sudo journalctl -b
http://0x0.st/HLPI.txt

seth · 2023-08-15 19:40:24

"ip a" is missing but you've two equal (default) routers over two NICs connected to the same segment.
Is this a deliberate bonding attempt?
You're only running NM, so I'm not sure how you end up this way…

Jocomol · 2023-08-15 19:46:10

Pardon here is ip a: http://0x0.st/HLZN.txt

Not a bonding attempt but just my laptop being connected to wire and wifi and the same time. if it matters the issue occures regardless of what I am connected to.

seth · 2023-08-15 19:59:27

You don't want to have this kind of routing collision.
Please cause this to happen w/ only the wifi connection and only onr default routing entry - something's off here.
The default behavior would be that NM establishes a wifi connection if the wired ones all lose the carrier (ie. you plug the cable)

An important element may be that you're apparently using a USB ethernet dongle, which might drop out frequently because of https://wiki.archlinux.org/title/Power_ … utosuspend

Jocomol · 2023-08-16 17:28:35

I've just tested it just with wifi, it still happens. Should I post any logs?

seth · 2023-08-16 19:07:54

Yes (system journal), also did you check "ip a; ip r" when that happened?

Jocomol · 2023-08-17 09:41:46

ip a: http://0x0.st/HLSg.txt
ip r: http://0x0.st/HLSE.txt
journalctl -b: http://0x0.st/HLS6.txt

Sorry wasn't able to reply yesterday. Same setup, same error today though. ip a and ip r looks the same. I was told by a friend that I should try installing the LTS kernel and I'll try that this evening.

seth · 2023-08-17 12:54:00

Nothing there suggests any connection issues - are other devices connected to the same (I assume it's a consumer grade) router/modem/switch combo?
Do you still get the

;; communications error to 8.8.8.8#53: timed out

from "dig @8.8.8.8" when this happens?

Koatao · 2023-08-17 13:31:35

Hello,

I noticed in the logs and in the output of systemctl that you are using bind9 (named.service) as a internal DNS server for your endpoint. Is it involved in the DNS resolution?

Could you post NM configuration and the content of /etc/resolv.conf?
I think we are missing something or a piece of information.

EDIT: By the way, mullvad is still running and seems to do something on every boot. Could you try to stop it completely?

Last edited by Koatao (2023-08-17 13:35:10)

seth · 2023-08-17 13:34:28

I'd be super-worried about the local resolver if it wasn't for the 8.8.8.8 timeout.
That's not explainable through a resolver/DNS issue.

Koatao · 2023-08-17 13:38:56

Mullvad daemon logs are concerning too. For instance:

Aug 17 09:25:31 brenodor mullvad-daemon[389]: [2023-08-17 09:25:31.857][mullvad_daemon::version][INFO] Starting mullvad-daemon - 2023.4 2023-06-27
Aug 17 09:25:31 brenodor mullvad-daemon[389]: [2023-08-17 09:25:31.865][mullvad_daemon::settings][INFO] Loading settings from /etc/mullvad-vpn/settings.json
Aug 17 09:25:31 brenodor mullvad-daemon[389]: [2023-08-17 09:25:31.874][mullvad_daemon::early_boot_firewall][INFO] Applying firewall policy Blocked. Blocking LAN. Allowing endpoint: none
Aug 17 09:25:31 brenodor mullvad-daemon[389]: [2023-08-17 09:25:31.874][talpid_core::firewall][INFO] Applying firewall policy: Blocked. Blocking LAN. Allowing endpoint: none

EDIT: looking more closely at the logs, mullvad is definitely applying its network and firewall configuration to the endpoint. I mean, it is popping everywhere in the logs!

Last edited by Koatao (2023-08-17 13:44:52)

seth · 2023-08-17 14:23:13

Yes, but https://bbs.archlinux.org/viewtopic.php … 1#p2115541

That being said, it that recollection might be hazy: lower the MTU

Last edited by seth (2023-08-17 14:42:26)

Jocomol · 2023-08-17 19:14:00

Yes I still get the timed out error. Yes other devices on the network don't have this issue. The concerning laptop also didn't have these issues when Windows was still installed from what I could tell. I also have these issues in other networks so we can really rule out any network errors.

I've disabled mullvad now completely.

>I noticed in the logs and in the output of systemctl that you are using bind9 (named.service) as a internal DNS server for your endpoint. Is it involved in the DNS resolution?

I don't know, how could I check?

resolv.conf: http://0x0.st/HLj0.txt
/etc/NetworkManager has just a bunch of folder, most of them empty except system-connections. Network-Manager.conf only got two comment lines.

[root@brenodor system-connections]# ls
Megafauna.nmconnection 'Wired connection 1.nmconnection' 'Wired connection 2.nmconnection'

Megafauna.nmconnection: http://0x0.st/HLjd.txt
Wired Connection 1: http://0x0.st/HLjn.txt
Wired Connection 2: http://0x0.st/HLj5.txt

Edit: how can I do code tags?

Last edited by Jocomol (2023-08-17 19:15:32)

seth · 2023-08-17 19:54:32

https://bbs.archlinux.org/help.php#bbcode

On the next occasion, try to "ping _gateway"
If this tries to ping 192.168.1.1, but fails, you lost contact to the router.
Maybe https://wiki.archlinux.org/title/Power_ … interfaces ?

If it cannot resolve 192.168.1.1 there's another issue.

Jocomol · 2023-08-18 09:26:23

I've tried and ping _gateway kept on working even without a noticeable drop in the time taken while dig timed out on 8.8.8.8. I've also checked my router but no problems there, also the same network is currently used by other devices which do not experience this issue (one of them being a similarly set up arch linux). Personally I suspect some kind of missing driver as I didn't configure much network wise and I noticed no problems on the previously installed windows 11.

Last edited by Jocomol (2023-08-18 09:30:18)

seth · 2023-08-18 10:21:43

Lower the MTU (try 1280, you'll have to do this for every NIC) - if that "fixes" it, it's most likely related to your VPN (despite all assurances)
If i doesn't, it's your router.

For your wifi:

sudo ip link set dev wlp61s0 mtu 1280

https://wiki.archlinux.org/title/Networ … eue_length on how to make that permanent.

Jocomol · 2023-08-25 15:17:06

Over the last week I've installed a seperate DNS and it looks like it solved the issue. Funnily enough the issue also doesn't occur anymore in other network. I guess the firmware upgrade I did solved the issue.

Arch Linux

#1 2023-08-15 17:00:57

[Solved] Random DNS Errors

#2 2023-08-15 18:25:53

Re: [Solved] Random DNS Errors

#3 2023-08-15 18:38:35

Re: [Solved] Random DNS Errors

#4 2023-08-15 18:44:04

Re: [Solved] Random DNS Errors

#5 2023-08-15 18:47:34

Re: [Solved] Random DNS Errors

#6 2023-08-15 18:53:12

Re: [Solved] Random DNS Errors

#7 2023-08-15 19:04:00

Re: [Solved] Random DNS Errors

#8 2023-08-15 19:06:37

Re: [Solved] Random DNS Errors

#9 2023-08-15 19:40:24

Re: [Solved] Random DNS Errors

#10 2023-08-15 19:46:10

Re: [Solved] Random DNS Errors

#11 2023-08-15 19:59:27

Re: [Solved] Random DNS Errors

#12 2023-08-16 17:28:35

Re: [Solved] Random DNS Errors

#13 2023-08-16 19:07:54

Re: [Solved] Random DNS Errors

#14 2023-08-17 09:41:46

Re: [Solved] Random DNS Errors

#15 2023-08-17 12:54:00

Re: [Solved] Random DNS Errors

#16 2023-08-17 13:31:35

Re: [Solved] Random DNS Errors

#17 2023-08-17 13:34:28

Re: [Solved] Random DNS Errors

#18 2023-08-17 13:38:56

Re: [Solved] Random DNS Errors

#19 2023-08-17 14:23:13

Re: [Solved] Random DNS Errors

#20 2023-08-17 19:14:00

Re: [Solved] Random DNS Errors

#21 2023-08-17 19:54:32

Re: [Solved] Random DNS Errors

#22 2023-08-18 09:26:23

Re: [Solved] Random DNS Errors

#23 2023-08-18 10:21:43

Re: [Solved] Random DNS Errors

#24 2023-08-25 15:17:06

Re: [Solved] Random DNS Errors

Board footer