SVI/netbird
1
0
Fork 0
mirror of https://github.com/netbirdio/netbird.git synced 2026-05-31 04:22:11 -04:00
Code Issues 3.0k Packages Projects Releases 10 Wiki Activity

Unable to connect via P2P even public IPv4 is available #1224

New Issue
Open
opened 2025-11-20 05:26:19 -05:00 by saavagebueno · 7 comments
saavagebueno commented 2025-11-20 05:26:19 -05:00
Owner
Copy Link

Originally created by @Integral-Tech on GitHub (Sep 9, 2024).

Describe the problem

I am pretty sure that one of my peers has public IPv4 address, however I fail to connect two peers via P2P.

  • Peer 1: Arch Linux (without public IPv4)
  • Peer 2: OpenWrt (with public IPv4)

To Reproduce

Steps to reproduce the behavior:

  1. Start Netbird service on two peers (one of them has public IPv4 address)
  2. Run netbird status -d

Expected behavior

  Status: Connected
  -- detail --
  Connection type: P2P
  Direct: true

Are you using NetBird Cloud?

Yes

NetBird version

0.28.9

NetBird status -dA output:

Peers detail:
 *.netbird.cloud:
  NetBird IP: *
  Public key: *
  Status: Connected
  -- detail --
  Connection type: Relayed
  Direct: true
  ICE candidate (Local/Remote): host/relay
  ICE candidate endpoints (Local/Remote): *
  Last connection update: 14 seconds ago
  Last WireGuard handshake: -
  Transfer status (received/sent) 444 B/424 B
  Quantum resistance: false
  Routes: 192.168.0.0/16
  Latency: 426.126669ms

OS: linux/amd64
Daemon version: 0.28.9
CLI version: 0.28.9
Management: Connected to https://api.netbird.io:443
Signal: Connected to https://signal.netbird.io:443
Relays: 
  [stun:stun.netbird.io:5555] is Available
  [turns:turn.netbird.io:443?transport=tcp] is Available
Nameservers: 
FQDN: *.netbird.cloud
NetBird IP: *
Interface type: Kernel
Quantum resistance: false
Routes: -
Peers count: 1/1 Connected
Originally created by @Integral-Tech on GitHub (Sep 9, 2024). **Describe the problem** I am pretty sure that one of my peers has public IPv4 address, however I fail to connect two peers via P2P. - Peer 1: Arch Linux (without public IPv4) - Peer 2: OpenWrt (with public IPv4) **To Reproduce** Steps to reproduce the behavior: 1. Start Netbird service on two peers (one of them has public IPv4 address) 2. Run ```netbird status -d``` **Expected behavior** ``` Status: Connected -- detail -- Connection type: P2P Direct: true ``` **Are you using NetBird Cloud?** Yes **NetBird version** 0.28.9 **NetBird status -dA output:** ``` Peers detail: *.netbird.cloud: NetBird IP: * Public key: * Status: Connected -- detail -- Connection type: Relayed Direct: true ICE candidate (Local/Remote): host/relay ICE candidate endpoints (Local/Remote): * Last connection update: 14 seconds ago Last WireGuard handshake: - Transfer status (received/sent) 444 B/424 B Quantum resistance: false Routes: 192.168.0.0/16 Latency: 426.126669ms OS: linux/amd64 Daemon version: 0.28.9 CLI version: 0.28.9 Management: Connected to https://api.netbird.io:443 Signal: Connected to https://signal.netbird.io:443 Relays: [stun:stun.netbird.io:5555] is Available [turns:turn.netbird.io:443?transport=tcp] is Available Nameservers: FQDN: *.netbird.cloud NetBird IP: * Interface type: Kernel Quantum resistance: false Routes: - Peers count: 1/1 Connected ```
saavagebueno added the bugtriage-needed labels 2025-11-20 05:26:19 -05:00
saavagebueno commented 2025-11-20 05:26:20 -05:00
Author
Owner
Copy Link

@pascal-fischer commented on GitHub (Sep 9, 2024):

Hi @Integral-Tech,
to have a full P2P connection both peers need to be reachable directly. In your status output you see
ICE candidate (Local/Remote): host/relay so one way is connected on host level (due to the public IP) which would allow P2P but the opposite direction is relayed. Thats why overall the connection type is listed as Relayed

@pascal-fischer commented on GitHub (Sep 9, 2024): Hi @Integral-Tech, to have a full P2P connection both peers need to be reachable directly. In your status output you see `ICE candidate (Local/Remote): host/relay` so one way is connected on host level (due to the public IP) which would allow P2P but the opposite direction is relayed. Thats why overall the connection type is listed as Relayed
saavagebueno commented 2025-11-20 05:26:20 -05:00
Author
Owner
Copy Link

@Marcus1Pierce commented on GitHub (Sep 10, 2024):

@Integral-Tech From what I know, if you want both peers to communicate via P2P, both peers must be on the same IP segment, or both peers must have a public IP and be accessible from outside.

#CMIIW

@Marcus1Pierce commented on GitHub (Sep 10, 2024): @Integral-Tech From what I know, if you want both peers to communicate via P2P, both peers must be on the same IP segment, or both peers must have a public IP and be accessible from outside. #CMIIW
saavagebueno commented 2025-11-20 05:26:20 -05:00
Author
Owner
Copy Link

@Spiritreader commented on GitHub (Oct 14, 2024):

Hi @Integral-Tech, to have a full P2P connection both peers need to be reachable directly. In your status output you see ICE candidate (Local/Remote): host/relay so one way is connected on host level (due to the public IP) which would allow P2P but the opposite direction is relayed. Thats why overall the connection type is listed as Relayed

If one direction can connect via host, isn't it possible to just establish a direct wireguard connection?
This is essentially the base scenario of manually setting up a wireguard tunnel for machines behind CGNAT:

  • Open whatever port (12345) on PeerA.
  • Configure wireguard on CTNATed PeerB to add PeerA with the address of PeerA:12345.

Now, as long as there is a keepalive interval to prevent the CGNAT's stateful FW from closing the connection, you're good to go.
I had to do this fairly often for peers before I used netbird, as some of my devices are behind CGNAT.

Having host/relay, kinda sounds like calling a friend on his phone via the number they gave you, they pick up and say "Hello" and in order to respond you send your grandma to his house via bike.

if we had a TCP connection on port 80 for CGNATed devices, some form of return traffic must be allowed directly because we need some form of traffic to flow between devices if they have a connection to each other.

Or am I understanding sth fundamentally wrong here, for example that the ICE candidate doesn't show an established connection, but rather only the peer addresses?

@Spiritreader commented on GitHub (Oct 14, 2024): > Hi @Integral-Tech, to have a full P2P connection both peers need to be reachable directly. In your status output you see `ICE candidate (Local/Remote): host/relay` so one way is connected on host level (due to the public IP) which would allow P2P but the opposite direction is relayed. Thats why overall the connection type is listed as Relayed If one direction can connect via host, isn't it possible to just establish a direct wireguard connection? This is essentially the base scenario of manually setting up a wireguard tunnel for machines behind CGNAT: - Open whatever port (12345) on PeerA. - Configure wireguard on CTNATed PeerB to add PeerA with the address of PeerA:12345. Now, as long as there is a keepalive interval to prevent the CGNAT's stateful FW from closing the connection, you're good to go. I had to do this fairly often for peers before I used netbird, as some of my devices are behind CGNAT. Having host/relay, kinda sounds like calling a friend on his phone via the number they gave you, they pick up and say "Hello" and in order to respond you send your grandma to his house via bike. if we had a TCP connection on port 80 for CGNATed devices, some form of return traffic must be allowed *directly* because we need some form of traffic to flow between devices if they have a connection to each other. Or am I understanding sth fundamentally wrong here, for example that the ICE candidate doesn't show an established connection, but rather only the peer addresses?
saavagebueno commented 2025-11-20 05:26:20 -05:00
Author
Owner
Copy Link

@vampywiz17 commented on GitHub (Oct 25, 2024):

Hi @Integral-Tech, to have a full P2P connection both peers need to be reachable directly. In your status output you see ICE candidate (Local/Remote): host/relay so one way is connected on host level (due to the public IP) which would allow P2P but the opposite direction is relayed. Thats why overall the connection type is listed as Relayed

If one direction can connect via host, isn't it possible to just establish a direct wireguard connection? This is essentially the base scenario of manually setting up a wireguard tunnel for machines behind CGNAT:

  • Open whatever port (12345) on PeerA.
  • Configure wireguard on CTNATed PeerB to add PeerA with the address of PeerA:12345.

Now, as long as there is a keepalive interval to prevent the CGNAT's stateful FW from closing the connection, you're good to go. I had to do this fairly often for peers before I used netbird, as some of my devices are behind CGNAT.

Having host/relay, kinda sounds like calling a friend on his phone via the number they gave you, they pick up and say "Hello" and in order to respond you send your grandma to his house via bike.

if we had a TCP connection on port 80 for CGNATed devices, some form of return traffic must be allowed directly because we need some form of traffic to flow between devices if they have a connection to each other.

Or am I understanding sth fundamentally wrong here, for example that the ICE candidate doesn't show an established connection, but rather only the peer addresses?

This is how it work Tailscale. not need to "active" both client, it enought that only one is active.

@vampywiz17 commented on GitHub (Oct 25, 2024): > > Hi @Integral-Tech, to have a full P2P connection both peers need to be reachable directly. In your status output you see `ICE candidate (Local/Remote): host/relay` so one way is connected on host level (due to the public IP) which would allow P2P but the opposite direction is relayed. Thats why overall the connection type is listed as Relayed > > If one direction can connect via host, isn't it possible to just establish a direct wireguard connection? This is essentially the base scenario of manually setting up a wireguard tunnel for machines behind CGNAT: > > * Open whatever port (12345) on PeerA. > * Configure wireguard on CTNATed PeerB to add PeerA with the address of PeerA:12345. > > Now, as long as there is a keepalive interval to prevent the CGNAT's stateful FW from closing the connection, you're good to go. I had to do this fairly often for peers before I used netbird, as some of my devices are behind CGNAT. > > Having host/relay, kinda sounds like calling a friend on his phone via the number they gave you, they pick up and say "Hello" and in order to respond you send your grandma to his house via bike. > > if we had a TCP connection on port 80 for CGNATed devices, some form of return traffic must be allowed _directly_ because we need some form of traffic to flow between devices if they have a connection to each other. > > Or am I understanding sth fundamentally wrong here, for example that the ICE candidate doesn't show an established connection, but rather only the peer addresses? This is how it work Tailscale. not need to "active" both client, it enought that only one is active.
saavagebueno commented 2025-11-20 05:26:20 -05:00
Author
Owner
Copy Link

@etranger7 commented on GitHub (Apr 16, 2025):

Hi, I have a similar problem. I am fairly certain my firewall configuration is correct.
In fact, before I reset my firewall, I had a P2P connection. Unless I'm missing something, I reinstated the same rules but this time I'm getting
ICE candidate (Local/Remote): srflx/relay
instead of a P2P connection.
@Integral-Tech were you able to figure this out?

@etranger7 commented on GitHub (Apr 16, 2025): Hi, I have a similar problem. I am fairly certain my firewall configuration is correct. In fact, before I reset my firewall, I had a P2P connection. Unless I'm missing something, I reinstated the same rules but this time I'm getting ICE candidate (Local/Remote): srflx/relay instead of a P2P connection. @Integral-Tech were you able to figure this out?
saavagebueno commented 2025-11-20 05:26:20 -05:00
Author
Owner
Copy Link

@nazarewk commented on GitHub (Apr 17, 2025):

@etranger7 could you send us a debug bundle (without anonymization -A) for ~5 minutes to review how the connectivity is being established between the peers? Also, please state which peers exactly those are.
You can mail it to support at nebird.io or send directly to me on Slack (kdn).

I have checked myself with one host (my laptop) running behind CGNAT (not my regular ISP) and the other one having a dedicated public IPv4 (Hetzner VM), and the connectivity is still established with P2P.

@nazarewk commented on GitHub (Apr 17, 2025): @etranger7 could you send us a debug bundle (without anonymization `-A`) for ~5 minutes to review how the connectivity is being established between the peers? Also, please state which peers exactly those are. You can mail it to support at nebird.io or send directly to me on Slack (kdn). I have checked myself with one host (my laptop) running behind CGNAT (not my regular ISP) and the other one having a dedicated public IPv4 (Hetzner VM), and the connectivity is still established with P2P.
saavagebueno commented 2025-11-20 05:26:21 -05:00
Author
Owner
Copy Link

@etranger7 commented on GitHub (Apr 20, 2025):

Hi, thanks for looking into this. Despite not changing anything in the firewall settings, the connection went back to P2P.
I forgot to mention, we're testing this self-hosted.
There seems to be some confusion over how peers make connections to the server and other peers when they are behind a firewall.
I went through the documentation and I understand the general principles of signaling and ICE.
However, I don't quite understand some specifics.

For example. Let's assume we have a peer behind a firewall. The firewall allows all outbound connections. There are no specific NAT rules pointing to this peer.
This peer should still be able to make P2P connections to other peers outside of that LAN by making an outbound connection, to both the management server and the other peer, isn't it?

@etranger7 commented on GitHub (Apr 20, 2025): Hi, thanks for looking into this. Despite not changing anything in the firewall settings, the connection went back to P2P. I forgot to mention, we're testing this self-hosted. There seems to be some confusion over how peers make connections to the server and other peers when they are behind a firewall. I went through the documentation and I understand the general principles of signaling and ICE. However, I don't quite understand some specifics. For example. Let's assume we have a peer behind a firewall. The firewall allows all outbound connections. There are no specific NAT rules pointing to this peer. This peer should still be able to make P2P connections to other peers outside of that LAN by making an outbound connection, to both the management server and the other peer, isn't it?
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
2021 Q4
2022 Q1
2022 Q1
accessibility
acl
agent
agent
Android
Android
api
authentik
automation
azure
battery-usage
bug
cache
client
client-ui
cloud
cloud-only
cloudflare
community
compatibility
config-idp
config-issue
connection
contribution
coturn
cross-vpn
dashboard
data-usage
distribution
dns
docker
documentation
duplicate
enhancement
enhancement
event-stream
feature-request
freebsd
getting-started
go
good first issue
gui
help wanted
home-assistant
idp
inconsistency
integration
integrations
ios
ipv6
jwt
k8s
keycloak
linux
login
macos
management-service
missing-docs
mobile
moved-internal
needs-review
netbird-ui
networking
new-platform
nginx
notification
okta
openwrt
packaging
peer-management
peer-management
peer-management
performance
postgres
posture-checks
psk
pull-request
Mirrored from GitHub Pull Request
 question
refactor
relay
release
rfc
routes
security
security-related
self-hosting
server
signal
sleep-issue
ssh
ssl
status
store
synology
system-compatibility-issue
test-suite
third-party-integration
triage
triage-needed
troubleshooting
UX
waiting-feedback
windows
wontfix
zitadel
No Label bug triage-needed
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
saavagebueno
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: SVI/netbird#1224
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?