search

srsran / srsRAN_Project

Open source O-RAN 5G CU/DU solution from Software Radio Systems (SRS) https://docs.srsran.com/projects/project
https://www.srsran.com
GNU Affero General Public License v3.0
491 stars 166 forks source link

UEs Connected but can't reach the core #613

Closed mas-oliveira closed 1 month ago

mas-oliveira commented 5 months ago

Issue Description

Hello, I connected 3 UEs to the gNB, now I was trying to ping the core through the UEs but I can't.

Setup Details

I have the O-RAN SC RIC and the 5G Core running on one machine and the UEs + gNB are running on another machine.

The PRACH detector will not meet the performance requirements with the configuration {Format 0, ZCZ 0, SCS 1.25kHz, Rx ports 1}. Lower PHY in executor blocking mode.

--== srsRAN gNB (commit 1483bda30) ==--

Connecting to AMF on 10.53.1.2:38412 Available radio types: zmq. Connecting to NearRT-RIC on 10.0.2.10:36421 Cell pci=1, bw=20 MHz, 1T1R, dl_arfcn=368500 (n3), dl_freq=1842.5 MHz, dl_ssb_arfcn=368410, ul_freq=1747.5 MHz

==== gNodeB started === Type to view trace

The gNB can connect to the core and the RIC correctly. And all the UEs can connect to the gNB:

Opening 1 channels in RF device=zmq with args=tx_port=tcp://127.0.0.1:2101,rx_port=tcp://127.0.0.1:2100,base_srate=23.04e6 Supported RF device list: zmq file CHx base_srate=23.04e6 Current sample rate is 1.92 MHz with a base rate of 23.04 MHz (x12 decimation) CH0 rx_port=tcp://127.0.0.1:2100 CH0 tx_port=tcp://127.0.0.1:2101 Current sample rate is 23.04 MHz with a base rate of 23.04 MHz (x1 decimation) Current sample rate is 23.04 MHz with a base rate of 23.04 MHz (x1 decimation) Waiting PHY to initialize ... done! Attaching UE... Random Access Transmission: prach_occasion=0, preamble_index=55, ra-rnti=0x39, tti=174 Random Access Complete. c-rnti=0x4602, ta=0 RRC Connected PDU Session Establishment successful. IP: 10.45.1.2 RRC NR reconfiguration successful.

These are the routes of the UEs

sudo ip netns exec ue1 ip route show default via 10.45.1.1 dev tun_srsue 10.45.1.0/24 dev tun_srsue proto kernel scope link src 10.45.1.2

And these are the interfaces available on the machine with the gNB + UEs

marco@darlene-G5-KC:~/srsRAN_Project/build/apps/gnb$ ifconfig br-b84c28fd1d11: flags=4099<UP,BROADCAST,MULTICAST> mtu 1500 inet 10.0.2.1 netmask 255.255.255.0 broadcast 10.0.2.255 ether 02:42:d0:92:33:65 txqueuelen 0 (Ethernet) RX packets 0 bytes 0 (0.0 B) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 0 bytes 0 (0.0 B) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

docker0: flags=4099<UP,BROADCAST,MULTICAST> mtu 1500 inet 172.17.0.1 netmask 255.255.0.0 broadcast 172.17.255.255 ether 02:42:bb:11:00:0a txqueuelen 0 (Ethernet) RX packets 0 bytes 0 (0.0 B) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 0 bytes 0 (0.0 B) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

enp8s0f1: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500 inet 10.42.0.130 netmask 255.255.255.0 broadcast 10.42.0.255 inet6 fe80::6721:e0e2:8fc7:ae4c prefixlen 64 scopeid 0x20 ether 80:fa:5b:96:c5:1d txqueuelen 1000 (Ethernet) RX packets 1473 bytes 280509 (280.5 KB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 6332 bytes 539234 (539.2 KB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

enx3a65b2cbdb60: flags=4099<UP,BROADCAST,MULTICAST> mtu 1500 ether 3a:65:b2:cb:db:60 txqueuelen 1000 (Ethernet) RX packets 0 bytes 0 (0.0 B) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 0 bytes 0 (0.0 B) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

lo: flags=73<UP,LOOPBACK,RUNNING> mtu 65536 inet 127.0.0.1 netmask 255.0.0.0 inet6 ::1 prefixlen 128 scopeid 0x10 loop txqueuelen 1000 (Local Loopback) RX packets 29082748 bytes 357893898753 (357.8 GB) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 29082748 bytes 357893898753 (357.8 GB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

wlp7s0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500 inet 10.43.78.214 netmask 255.255.240.0 broadcast 10.43.79.255 inet6 fe80::3654:27a4:b4e9:554f prefixlen 64 scopeid 0x20 ether 08:f8:bc:67:05:74 txqueuelen 1000 (Ethernet) RX packets 15188 bytes 16668777 (16.6 MB) RX errors 0 dropped 361 overruns 0 frame 0 TX packets 10411 bytes 2069441 (2.0 MB) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

Shouldn't be here a dev tun_srsue interface?

EDIT I added the ogstun interface on the machine with gNB + UEs: ogstun: flags=4241<UP,POINTOPOINT,NOARP,MULTICAST> mtu 1500 unspec 00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00 txqueuelen 500 (UNSPEC) RX packets 0 bytes 0 (0.0 B) RX errors 0 dropped 0 overruns 0 frame 0 TX packets 0 bytes 0 (0.0 B) TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

And I have the following route table on UE1: $ sudo ip netns exec ue1 route -n Kernel IP routing table Destination Gateway Genmask Flags Metric Ref Use Iface 0.0.0.0 10.45.1.1 0.0.0.0 UG 0 0 0 tun_srsue 10.45.1.0 0.0.0.0 255.255.255.0 U 0 0 0 tun_srsue

But I still can't reach the core $ sudo ip netns exec ue1 ping 10.45.1.1 PING 10.45.1.1 (10.45.1.1) 56(84) bytes of data. ^C --- 10.45.1.1 ping statistics --- 3 packets transmitted, 0 received, 100% packet loss, time 2043ms

Should I add a rule where the packet that hit the ogstun interface should be forward to the interface where is connected the computer with the core and the RIC?

Expected Behavior

While performing the sudo ip netns exec ue1 ping 10.45.1.1 command the UE was able to reach the core network.

Actual Behaviour

marco@darlene-G5-KC:~/Downloads$ sudo ip netns exec ue1 ping 10.45.1.1 PING 10.45.1.1 (10.45.1.1) 56(84) bytes of data. ^C --- 10.45.1.1 ping statistics --- 4 packets transmitted, 0 received, 100% packet loss, time 3072ms

pgawlowicz commented 5 months ago

Do you run the dockerized version of the Open5gs or one installed manually?

mas-oliveira commented 5 months ago

Hello Piotr, I'm running the dockerized version available on srsRAN_Project/docker on the NuC. But I don't know if the UE ping is hitting the core.

pgawlowicz commented 5 months ago

could you show routing tables on the PC runnning gnb and open5gs container?

mas-oliveira commented 5 months ago

I'm executing tcpdump to capture the ICMP packets on the core machine. When I start the ping using the UE1 (running on a NuC) I capture the following packets: 13:03:32.452547 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136 13:03:32.452549 veth1eeab61 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136 13:03:33.478944 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136 13:03:33.478947 veth1eeab61 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136 13:03:34.473379 lo In IP localhost > localhost: ICMP localhost udp port 8805 unreachable, length 66 13:03:34.519852 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136 13:03:34.519854 veth1eeab61 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136

mas-oliveira commented 5 months ago

Core computer:

Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
0.0.0.0         10.3.10.1       0.0.0.0         UG    600    0        0 wlp58s0
10.0.2.0        0.0.0.0         255.255.255.0   U     0      0        0 br-0ae98bbb2186
10.3.10.0       0.0.0.0         255.255.255.0   U     600    0        0 wlp58s0
10.42.0.0       0.0.0.0         255.255.255.0   U     100    0        0 eno1
10.45.0.0       0.0.0.0         255.255.0.0     U     0      0        0 ogstun
10.53.1.0       0.0.0.0         255.255.255.0   U     0      0        0 br-26141c6f72a5
169.254.0.0     0.0.0.0         255.255.0.0     U     1000   0        0 eno1
172.17.0.0      0.0.0.0         255.255.0.0     U     0      0        0 docker0
172.19.1.0      0.0.0.0         255.255.255.0   U     0      0        0 br-60b9fad393f0

gNB + UEs computer:

Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
0.0.0.0         10.42.0.1       0.0.0.0         UG    100    0        0 enp8s0f1
10.42.0.0       0.0.0.0         255.255.255.0   U     100    0        0 enp8s0f1
172.17.0.0      0.0.0.0         255.255.0.0     U     0      0        0 docker0
pgawlowicz commented 5 months ago

ok, so you do not need to add ogstun manually on the core network PC. Please remove it.

Instead, add the following route on the PC running open5gs container:

sudo ip ro add 10.45.0.0/16 via 10.53.1.2

Then, I am not sure how the gnb PC is able to reach AMF if it has no route to 10.53.1.2 (does it work?) You need to add this route and probably enable IP forwarding on the gnb pc.

mas-oliveira commented 5 months ago

I have now the following routes on the core computer:

marco@marco-NUC7i7BNH:~/Desktop$ ip route show
default via 10.3.10.1 dev wlp58s0 proto dhcp metric 600 
10.0.2.0/24 dev br-0ae98bbb2186 proto kernel scope link src 10.0.2.1 
10.3.10.0/24 dev wlp58s0 proto kernel scope link src 10.3.10.118 metric 600 
10.42.0.0/24 dev eno1 proto kernel scope link src 10.42.0.1 metric 100 
10.45.0.0/16 via 10.53.1.2 dev br-26141c6f72a5 
10.53.1.0/24 dev br-26141c6f72a5 proto kernel scope link src 10.53.1.1 
169.254.0.0/16 dev eno1 scope link metric 1000 
172.17.0.0/16 dev docker0 proto kernel scope link src 172.17.0.1 linkdown 
172.19.1.0/24 dev br-60b9fad393f0 proto kernel scope link src 172.19.1.1

But then, when I try to ping 10.45.1.1 using UE1 I got the following output:

$ sudo tcpdump -i any icmp
13:22:39.289176 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
13:22:39.289179 vethbcb83d5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
13:22:40.355509 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
13:22:40.355511 vethbcb83d5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
13:22:41.303954 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136

About the gNB computer, since when I connect the gNB I have the following output:

$ sudo ./gnb -c gnb_zmq.yaml 

The PRACH detector will not meet the performance requirements with the configuration {Format 0, ZCZ 0, SCS 1.25kHz, Rx ports 1}.
Lower PHY in executor blocking mode.

--== srsRAN gNB (commit 1483bda30) ==--

Connecting to AMF on 10.53.1.2:38412
Available radio types: zmq.
Connecting to NearRT-RIC on 10.0.2.10:36421
Cell pci=1, bw=20 MHz, 1T1R, dl_arfcn=368500 (n3), dl_freq=1842.5 MHz, dl_ssb_arfcn=368410, ul_freq=1747.5 MHz

==== gNodeB started ===
Type <t> to view trace

Isn't this because I can reach AMF using the AMF computer or do I need to specify a route to it?

pgawlowicz commented 5 months ago

hmm did you enable IP forwardind on the gnb PC?

sudo sysctl -w net.ipv4.ip_forward=1
sudo iptables -t nat -A POSTROUTING -o <IFNAME> -j MASQUERADE
mas-oliveira commented 5 months ago

I have done this on the gNB computer:

marco@darlene-G5-KC:~/srsRAN_4G/build/srsue/src$ sudo sysctl -w net.ipv4.ip_forward=1
net.ipv4.ip_forward = 1

marco@darlene-G5-KC:~/srsRAN_4G/build/srsue/src$ sudo iptables -t nat -A POSTROUTING -o wlp7s0 -j MASQUERADE

And now I have the following route table:

marco@darlene-G5-KC:~/srsRAN_4G/build/srsue/src$ route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
0.0.0.0         10.42.0.1       0.0.0.0         UG    100    0        0 enp8s0f1
0.0.0.0         10.3.10.1       0.0.0.0         UG    600    0        0 wlp7s0
10.3.10.0       0.0.0.0         255.255.255.0   U     600    0        0 wlp7s0
10.42.0.0       0.0.0.0         255.255.255.0   U     100    0        0 enp8s0f1
169.254.0.0     0.0.0.0         255.255.0.0     U     1000   0        0 wlp7s0
172.17.0.0      0.0.0.0         255.255.0.0     U     0      0        0 docker0

The gNB and the UEs can connect but the ping still not working

marco@darlene-G5-KC:~/srsRAN_4G/build/srsue/src$ sudo ip netns exec ue1 ping 10.45.1.1
PING 10.45.1.1 (10.45.1.1) 56(84) bytes of data.
^C
--- 10.45.1.1 ping statistics ---
10 packets transmitted, 0 received, 100% packet loss, time 9218ms

On the core computer while making the TCP Dump I still can check that I'm receiving something from the gNB computer when I start the ping on the UE1:

13:39:27.960045 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
13:39:27.960047 vethbcb83d5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
13:39:28.988822 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
13:39:28.988824 vethbcb83d5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
13:39:30.028386 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
13:39:30.028388 vethbcb83d5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136

But I still without getting the ICMP reply on the UE1

pgawlowicz commented 5 months ago

which PC is 10.42.0.130?

could you share your gnb and UE config?

mas-oliveira commented 5 months ago

gNB + UE Computer (NuC): 10.42.0.130 Core + SC RIC (Darlene): 10.42.0.1

confs.zip

pgawlowicz commented 5 months ago

could you check if you can ping 10.53.1.2 from gnb pc?

mas-oliveira commented 5 months ago 
marco@darlene-G5-KC:~/srsRAN_4G/build/srsue/src$ ping 10.53.1.2
PING 10.53.1.2 (10.53.1.2) 56(84) bytes of data.
64 bytes from 10.53.1.2: icmp_seq=1 ttl=63 time=0.596 ms
64 bytes from 10.53.1.2: icmp_seq=2 ttl=63 time=0.376 ms
64 bytes from 10.53.1.2: icmp_seq=3 ttl=63 time=0.797 ms
^C
--- 10.53.1.2 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2028ms
rtt min/avg/max/mdev = 0.376/0.589/0.797/0.171 ms
pgawlowicz commented 5 months ago

and if you can ping the gnb from the open5gs container?

docker exec -it open5gs_5gc bash
ping 10.53.1.1
mas-oliveira commented 5 months ago

Yes, I was also able to do it:

root@ec86af9c1326:/open5gs# ping 10.53.1.1
PING 10.53.1.1 (10.53.1.1) 56(84) bytes of data.
64 bytes from 10.53.1.1: icmp_seq=1 ttl=64 time=0.079 ms
64 bytes from 10.53.1.1: icmp_seq=2 ttl=64 time=0.033 ms
64 bytes from 10.53.1.1: icmp_seq=3 ttl=64 time=0.055 ms
64 bytes from 10.53.1.1: icmp_seq=4 ttl=64 time=0.041 ms
64 bytes from 10.53.1.1: icmp_seq=5 ttl=64 time=0.034 ms
^C
--- 10.53.1.1 ping statistics ---
5 packets transmitted, 5 received, 0% packet loss, time 4103ms
rtt min/avg/max/mdev = 0.033/0.048/0.079/0.017 ms
pgawlowicz commented 5 months ago

ah, my mistake, you need to check if you can ping the gnb (10.42.0.130) from open5gs container.

mas-oliveira commented 5 months ago

Ah ok, this one I'm not able to Ping

root@ec86af9c1326:/open5gs# ping 10.42.0.130
PING 10.42.0.130 (10.42.0.130) 56(84) bytes of data.
From 10.53.1.1 icmp_seq=1 Destination Port Unreachable
From 10.53.1.1 icmp_seq=2 Destination Port Unreachable
From 10.53.1.1 icmp_seq=3 Destination Port Unreachable
^C
--- 10.42.0.130 ping statistics ---
3 packets transmitted, 0 received, +3 errors, 100% packet loss, time 2046ms
pgawlowicz commented 5 months ago

could you print the routing table inside the open5gs container?

mas-oliveira commented 5 months ago

Yes, since the result was so long I'll attach here the routing table:

open5gs_route.txt

pgawlowicz commented 5 months ago

you need to add a route to gnb over eth0 interface.

mas-oliveira commented 5 months ago

Sorry it didn't work

root@ec86af9c1326:/open5gs# ping 10.42.0.130
PING 10.42.0.130 (10.42.0.130) 56(84) bytes of data.
From 10.53.1.1 icmp_seq=1 Destination Port Unreachable
From 10.53.1.1 icmp_seq=2 Destination Port Unreachable
From 10.53.1.1 icmp_seq=3 Destination Port Unreachable
^C
--- 10.42.0.130 ping statistics ---
3 packets transmitted, 0 received, +3 errors, 100% packet loss, time 2046ms

I also tried to add the IP (10.42.0.130) instead of the subnet and also tried to add a route over ogstun interface. Neither of those worked

pgawlowicz commented 5 months ago

could you show the routes you tried to add? and the routing table?

mas-oliveira commented 5 months ago 
root@ec86af9c1326:/open5gs# ip route add 10.42.0.130 dev eth0
root@ec86af9c1326:/open5gs# ping 10.42.0.130
PING 10.42.0.130 (10.42.0.130) 56(84) bytes of data.
From 10.53.1.2 icmp_seq=1 Destination Host Unreachable
From 10.53.1.2 icmp_seq=2 Destination Host Unreachable
From 10.53.1.2 icmp_seq=3 Destination Host Unreachable
^C
--- 10.42.0.130 ping statistics ---
4 packets transmitted, 0 received, +3 errors, 100% packet loss, time 3076ms
pipe 4
root@ec86af9c1326:/open5gs# ip route delete 10.42.0.130 
root@ec86af9c1326:/open5gs# ip route add 10.42.0.130 dev ogstun
root@ec86af9c1326:/open5gs# ping 10.42.0.130
PING 10.42.0.130 (10.42.0.130) 56(84) bytes of data.
^C
--- 10.42.0.130 ping statistics ---
8 packets transmitted, 0 received, 100% packet loss, time 7156ms

root@ec86af9c1326:/open5gs# ip route delete 10.42.0.130
root@ec86af9c1326:/open5gs# ip route add 10.42.0.0/24 dev eth0  
root@ec86af9c1326:/open5gs# ping 10.42.0.130
PING 10.42.0.130 (10.42.0.130) 56(84) bytes of data.
From 10.53.1.2 icmp_seq=1 Destination Host Unreachable
From 10.53.1.2 icmp_seq=2 Destination Host Unreachable
From 10.53.1.2 icmp_seq=3 Destination Host Unreachable
^C
--- 10.42.0.130 ping statistics ---
4 packets transmitted, 0 received, +3 errors, 100% packet loss, time 3079ms
pipe 4
root@ec86af9c1326:/open5gs# ip route delete 10.42.0.0/24
root@ec86af9c1326:/open5gs# ip route add 10.42.0.0/24 dev ogstun
root@ec86af9c1326:/open5gs# ping 10.42.0.130
PING 10.42.0.130 (10.42.0.130) 56(84) bytes of data.
^C
--- 10.42.0.130 ping statistics ---
6 packets transmitted, 0 received, 100% packet loss, time 5119ms

I was adding and removing this route, so the routing table was the same as the one that I'd sent you with this new temporary test route

pgawlowicz commented 5 months ago

could you try sudo ip ro add 10.42.0.0/24 via 10.53.1.1

on the gnb PC, please add sudo ip ro add 10.53.1.0/24 via 10.42.0.1 (but since 10.42.0.1 is the default gw, maybe this one is not needed)

mas-oliveira commented 5 months ago

I added the route in the gNB PC just to be sure:

marco@darlene-G5-KC:~/srsRAN_4G/build/srsue/src$ sudo ip ro add 10.53.1.0/24 via 10.42.0.1
[sudo] password for marco: 
marco@darlene-G5-KC:~/srsRAN_4G/build/srsue/src$ ip ro
default via 10.42.0.1 dev enp8s0f1 proto dhcp metric 100 
default via 10.3.10.1 dev wlp7s0 proto dhcp metric 600 
10.3.10.0/24 dev wlp7s0 proto kernel scope link src 10.3.10.114 metric 600 
10.42.0.0/24 dev enp8s0f1 proto kernel scope link src 10.42.0.130 metric 100 
10.53.1.0/24 via 10.42.0.1 dev enp8s0f1 
169.254.0.0/16 dev wlp7s0 scope link metric 1000 
172.17.0.0/16 dev docker0 proto kernel scope link src 172.17.0.1 linkdown

Then I tried to make the change on the Open5GS image:

root@ec86af9c1326:/open5gs# ip route delete 10.42.0.0/24
root@ec86af9c1326:/open5gs# ip ro add 10.42.0.0/24 via 10.53.1.1
root@ec86af9c1326:/open5gs# ping 10.42.0.130
PING 10.42.0.130 (10.42.0.130) 56(84) bytes of data.
From 10.53.1.1 icmp_seq=1 Destination Port Unreachable
From 10.53.1.1 icmp_seq=2 Destination Port Unreachable
From 10.53.1.1 icmp_seq=3 Destination Port Unreachable
^C
--- 10.42.0.130 ping statistics ---
3 packets transmitted, 0 received, +3 errors, 100% packet loss, time 2028ms

root@ec86af9c1326:/open5gs# ping 10.42.0.130
PING 10.42.0.130 (10.42.0.130) 56(84) bytes of data.
From 10.53.1.1 icmp_seq=1 Destination Port Unreachable
From 10.53.1.1 icmp_seq=2 Destination Port Unreachable
From 10.53.1.1 icmp_seq=3 Destination Port Unreachable
^C
--- 10.42.0.130 ping statistics ---
3 packets transmitted, 0 received, +3 errors, 100% packet loss, time 2030ms

root@ec86af9c1326:/open5gs# ip ro
default via 10.53.1.1 dev eth0 
10.42.0.0/24 via 10.53.1.1 dev eth0
pgawlowicz commented 5 months ago

did you enable the IP forwarding on the pc running open5gs?

mas-oliveira commented 5 months ago

Yes. By doing this: sudo sysctl -w net.ipv4.ip_forward=1 sudo iptables -t nat -A POSTROUTING -s 10.45.0.0/16 ! -o ogstun -j MASQUERADE

pgawlowicz commented 5 months ago

hmm, but there is no ogstun interface on this PC, or?

mas-oliveira commented 5 months ago

Yes, you're right. Probably the command was just ignored:

I can show to you that on the computer where the core is running there is not OGSTUN interface:


marco@marco-NUC7i7BNH:~/Desktop$ ifconfig
br-0ae98bbb2186: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 10.0.2.1  netmask 255.255.255.0  broadcast 10.0.2.255
        inet6 fe80::42:1fff:fed2:a166  prefixlen 64  scopeid 0x20<link>
        ether 02:42:1f:d2:a1:66  txqueuelen 0  (Ethernet)
        RX packets 668  bytes 47528 (47.5 KB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 729  bytes 81325 (81.3 KB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

br-26141c6f72a5: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 10.53.1.1  netmask 255.255.255.0  broadcast 10.53.1.255
        inet6 fe80::42:73ff:fed5:8e7f  prefixlen 64  scopeid 0x20<link>
        ether 02:42:73:d5:8e:7f  txqueuelen 0  (Ethernet)
        RX packets 2051  bytes 195104 (195.1 KB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 3607  bytes 348353 (348.3 KB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

br-60b9fad393f0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 172.19.1.1  netmask 255.255.255.0  broadcast 172.19.1.255
        inet6 fe80::42:3cff:fe75:58d5  prefixlen 64  scopeid 0x20<link>
        ether 02:42:3c:75:58:d5  txqueuelen 0  (Ethernet)
        RX packets 23  bytes 3420 (3.4 KB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 80  bytes 15289 (15.2 KB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

docker0: flags=4099<UP,BROADCAST,MULTICAST>  mtu 1500
        inet 172.17.0.1  netmask 255.255.0.0  broadcast 172.17.255.255
        ether 02:42:e3:11:05:38  txqueuelen 0  (Ethernet)
        RX packets 0  bytes 0 (0.0 B)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 0  bytes 0 (0.0 B)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

eno1: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 10.42.0.1  netmask 255.255.255.0  broadcast 10.42.0.255
        inet6 fe80::f2e9:74dd:fd:a8b8  prefixlen 64  scopeid 0x20<link>
        ether 94:c6:91:a8:6d:4a  txqueuelen 1000  (Ethernet)
        RX packets 18200  bytes 2528860 (2.5 MB)
        RX errors 0  dropped 10  overruns 0  frame 0
        TX packets 17274  bytes 7237553 (7.2 MB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0
        device interrupt 16  memory 0xdc300000-dc320000  

lo: flags=73<UP,LOOPBACK,RUNNING>  mtu 65536
        inet 127.0.0.1  netmask 255.0.0.0
        inet6 ::1  prefixlen 128  scopeid 0x10<host>
        loop  txqueuelen 1000  (Local Loopback)
        RX packets 256095  bytes 35787497 (35.7 MB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 256095  bytes 35787497 (35.7 MB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

veth134695a: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet6 fe80::70e9:31ff:fe15:fd1d  prefixlen 64  scopeid 0x20<link>
        ether 72:e9:31:15:fd:1d  txqueuelen 0  (Ethernet)
        RX packets 0  bytes 0 (0.0 B)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 63  bytes 6266 (6.2 KB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

veth2a94813: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet6 fe80::2c89:54ff:feb5:50b0  prefixlen 64  scopeid 0x20<link>
        ether 2e:89:54:b5:50:b0  txqueuelen 0  (Ethernet)
        RX packets 43196  bytes 9394806 (9.3 MB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 43276  bytes 6464275 (6.4 MB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

veth5fed9de: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet6 fe80::2494:72ff:fea4:3c9  prefixlen 64  scopeid 0x20<link>
        ether 26:94:72:a4:03:c9  txqueuelen 0  (Ethernet)
        RX packets 33  bytes 2179 (2.1 KB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 89  bytes 8354 (8.3 KB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

veth7edc6e3: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet6 fe80::44b2:ff:fece:9fd1  prefixlen 64  scopeid 0x20<link>
        ether 46:b2:00:ce:9f:d1  txqueuelen 0  (Ethernet)
        RX packets 43932  bytes 7055226 (7.0 MB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 84089  bytes 9997522 (9.9 MB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

vethbba5402: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet6 fe80::6cc2:60ff:febe:2559  prefixlen 64  scopeid 0x20<link>
        ether 6e:c2:60:be:25:59  txqueuelen 0  (Ethernet)
        RX packets 1524  bytes 98484 (98.4 KB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 1059  bytes 74327 (74.3 KB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

vethbcb83d5: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet6 fe80::50d1:bbff:feb4:1ff0  prefixlen 64  scopeid 0x20<link>
        ether 52:d1:bb:b4:1f:f0  txqueuelen 0  (Ethernet)
        RX packets 1392  bytes 146188 (146.1 KB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 2686  bytes 224528 (224.5 KB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

vethc6eefe6: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet6 fe80::8c5b:fbff:fe3f:c8f4  prefixlen 64  scopeid 0x20<link>
        ether 8e:5b:fb:3f:c8:f4  txqueuelen 0  (Ethernet)
        RX packets 196  bytes 13772 (13.7 KB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 280  bytes 23722 (23.7 KB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

vethd5ed35e: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet6 fe80::f8c5:d2ff:fe03:c4fa  prefixlen 64  scopeid 0x20<link>
        ether fa:c5:d2:03:c4:fa  txqueuelen 0  (Ethernet)
        RX packets 125381  bytes 16306674 (16.3 MB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 85855  bytes 16356985 (16.3 MB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

vethf3f140a: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet6 fe80::f082:2bff:fe19:e724  prefixlen 64  scopeid 0x20<link>
        ether f2:82:2b:19:e7:24  txqueuelen 0  (Ethernet)
        RX packets 23  bytes 3742 (3.7 KB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 121  bytes 19769 (19.7 KB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0

wlp58s0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
        inet 10.3.10.118  netmask 255.255.255.0  broadcast 10.3.10.255
        inet6 fe80::2aca:75c7:8117:658e  prefixlen 64  scopeid 0x20<link>
        ether 48:a4:72:d9:b6:ab  txqueuelen 1000  (Ethernet)
        RX packets 245264  bytes 224032903 (224.0 MB)
        RX errors 0  dropped 0  overruns 0  frame 0
        TX packets 108072  bytes 48929782 (48.9 MB)
        TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0
``
mas-oliveira commented 5 months ago

I also noticed that when I capture the packets on the core computer while trying to ping the gNB one:

root@ec86af9c1326:/open5gs# ping 10.42.0.130
PING 10.42.0.130 (10.42.0.130) 56(84) bytes of data.
From 10.53.1.1 icmp_seq=1 Destination Port Unreachable
From 10.53.1.1 icmp_seq=2 Destination Port Unreachable
From 10.53.1.1 icmp_seq=3 Destination Port Unreachable
From 10.53.1.1 icmp_seq=4 Destination Port Unreachable
From 10.53.1.1 icmp_seq=5 Destination Port Unreachable
From 10.53.1.1 icmp_seq=6 Destination Port Unreachable
From 10.53.1.1 icmp_seq=7 Destination Port Unreachable
From 10.53.1.1 icmp_seq=8 Destination Port Unreachable
From 10.53.1.1 icmp_seq=9 Destination Port Unreachable
From 10.53.1.1 icmp_seq=10 Destination Port Unreachable

18:00:28.797128 lo    In  IP localhost > localhost: ICMP localhost udp port 8805 unreachable, length 66
18:00:29.261438 vethbcb83d5 P   IP 10.53.1.2 > 10.42.0.130: ICMP echo request, id 28, seq 4, length 64
18:00:29.261438 br-26141c6f72a5 In  IP 10.53.1.2 > 10.42.0.130: ICMP echo request, id 28, seq 4, length 64
18:00:29.261476 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 protocol 1 port 45424 unreachable, length 92
18:00:29.261479 vethbcb83d5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 protocol 1 port 45424 unreachable, length 92
18:00:29.345951 lo    In  IP localhost > localhost: ICMP localhost udp port 8805 unreachable, length 66
18:00:30.285318 vethbcb83d5 P   IP 10.53.1.2 > 10.42.0.130: ICMP echo request, id 28, seq 5, length 64
18:00:30.285318 br-26141c6f72a5 In  IP 10.53.1.2 > 10.42.0.130: ICMP echo request, id 28, seq 5, length 64
18:00:30.285351 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 protocol 1 port 30226 unreachable, length 92
18:00:30.285353 vethbcb83d5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 protocol 1 port 30226 unreachable, length 92
18:00:31.299876 lo    In  IP localhost > localhost: ICMP localhost udp port 8805 unreachable, length 66
18:00:31.309534 vethbcb83d5 P   IP 10.53.1.2 > 10.42.0.130: ICMP echo request, id 28, seq 6, length 64
18:00:31.309534 br-26141c6f72a5 In  IP 10.53.1.2 > 10.42.0.130: ICMP echo request, id 28, seq 6, length 64
18:00:31.309578 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 protocol 1 port 58034 unreachable, length 92
18:00:31.309581 vethbcb83d5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 protocol 1 port 58034 unreachable, length 92
18:00:31.848625 lo    In  IP localhost > localhost: ICMP localhost udp port 8805 unreachable, length 66
18:00:32.333303 vethbcb83d5 P   IP 10.53.1.2 > 10.42.0.130: ICMP echo request, id 28, seq 7, length 64
18:00:32.333303 br-26141c6f72a5 In  IP 10.53.1.2 > 10.42.0.130: ICMP echo request, id 28, seq 7, length 64
18:00:32.333336 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 protocol 1 port 63316 unreachable, length 92
18:00:32.333338 vethbcb83d5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 protocol 1 port 63316 unreachable, length 92

that the packets are being sent to 10.42.0.130 by the 10.53.1.2 and shouldn't be any problem with this since the gNB computer can ping the IP 10.53.1.2

yc541 commented 5 months ago

I have been checking on this post since I have a similar issue when running the dockerized Open5GS on a separate machine from the gNB machine (UE works fine when Open5GS is on the gNB machine). What I did was slightly different, in the docker-compose.yml, I uncommented the ports under the 5gc section: _ports:

- "9999:9999/tcp"

# Uncomment port to use the 5gc from outside the docker network
   - "38412:38412/sctp"_

And in the gnb yml the amf IP is set to the host IP of the core machine rather than 10.53.1.2. gNB connects to the core correctly (gNB and core machines are in the same subnet), and similarly UE attaches successfully but can't ping the core. Perhaps this is something alternative to try (although it doesn't work with my set up)?

In terms of the ogstun, this should be an interface inside the container rather than the core's host machine.

pgawlowicz commented 5 months ago

@yc541 indeed this is a good idea, did you enable IP forwarding on the gnb pc?

yc541 commented 5 months ago

@pgawlowicz Yes ip forwarding is enabled on the gNb and core machines. I can ping the gNB IP from inside the core container (10.53.1.2) but not the other way round. I added a route at the gNb machine ip route add 10.53.1.0/24 via gNB IP, and I can ping 10.53.1.1 from the gNB machine but not 10.53.1.2.

@mas-oliveira Yes this is what I did, I assume 10.42.0.1 is the core host IP and 10.42.0.130 is the gNB host IP? Looks like something else is using the 2152 port on the gNB machine.

mas-oliveira commented 5 months ago

Yes, I fixed it. I thought that was something running on the gNB machine and I got confused.

Now I also can connect. I'm not using the RIC to simplify also. I have the same as you do. Using the addr of the machine with the core on the AMF and I've uncommented the port for sctp on docker-composer:

    ports:
      - "3000:3000/tcp"
      - "15356:15356"
      - "54558:54558" ### socket ping info
      - "54559:54559" ### socket udp iperf info
    # Uncomment port to use the 5gc from outside the docker network
      - "38412:38412/sctp"

But it stills not to work. From my point of view what's happening now is that the UE packets aren't leaving from the gNB computer.

gNB computer:

marco@darlene-G5-KC:~/Downloads$ sudo ip netns exec ue1 ping 10.45.1.1
PING 10.45.1.1 (10.45.1.1) 56(84) bytes of data.
^C
--- 10.45.1.1 ping statistics ---
20 packets transmitted, 0 received, 100% packet loss, time 19463ms

Core computer:

marco@marco-NUC7i7BNH:~/Desktop$ sudo tcpdump -i eno1 icmp
tcpdump: verbose output suppressed, use -v[v]... for full protocol decode
listening on eno1, link-type EN10MB (Ethernet), snapshot length 262144 bytes
^C
0 packets captured
0 packets received by filter
0 packets dropped by kernel

Maybe this is because there isn't any link between the interface of the UE and a rule set on the the gNB machine to deal with this traffic marco@darlene-G5-KC:~/Downloads$ sudo ip netns exec ue1 ip ro 10.45.1.0/24 dev tun_srsue proto kernel scope link src 10.45.1.2

Am I wrong?

Edit:

marco@darlene-G5-KC:~/Downloads$ ping 10.53.1.1
PING 10.53.1.1 (10.53.1.1) 56(84) bytes of data.
64 bytes from 10.53.1.1: icmp_seq=1 ttl=64 time=0.023 ms
64 bytes from 10.53.1.1: icmp_seq=2 ttl=64 time=0.028 ms
64 bytes from 10.53.1.1: icmp_seq=3 ttl=64 time=0.025 ms
^C
--- 10.53.1.1 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2031ms
rtt min/avg/max/mdev = 0.023/0.025/0.028/0.002 ms
marco@darlene-G5-KC:~/Downloads$ ping 10.53.1.2
PING 10.53.1.2 (10.53.1.2) 56(84) bytes of data.
From 10.43.64.1 icmp_seq=1 Destination Port Unreachable
From 10.43.64.1 icmp_seq=2 Destination Port Unreachable
^C
--- 10.53.1.2 ping statistics ---
2 packets transmitted, 0 received, +2 errors, 100% packet loss, time 1001ms

I have the same issue as @yc541 I can also ping 10.53.1.1 but not 10.53.1.2

pgawlowicz commented 5 months ago

the RTT of the ping to 10.53.1.1 seems to be quite low, are you sure it is not pinging local interface of the gnb pc?

mas-oliveira commented 5 months ago

You're right. I can't ping 10.53.1.2 and when I try to I can't even see the packets arriving at the core computer. On the gNB computer:

default via 10.43.64.1 dev wlp7s0 
default via 10.43.64.1 dev wlp7s0 proto dhcp metric 600 
default via 10.42.0.1 dev enp8s0f1 proto dhcp metric 20100 
10.0.2.0/24 dev enp8s0f1 proto kernel scope link src 10.42.0.130 metric 100 
10.42.0.0/24 dev enp8s0f1 proto kernel scope link src 10.42.0.130 metric 100 
10.43.64.0/20 dev wlp7s0 proto kernel scope link src 10.43.78.214 metric 600 
10.53.1.2 dev enp8s0f1 proto kernel scope link src 10.42.0.130 metric 100 
169.254.0.0/16 dev wlp7s0 scope link metric 1000 
172.17.0.0/16 dev docker0 proto kernel scope link src 172.17.0.1 linkdown

marco@darlene-G5-KC:~/Downloads$ ping 10.53.1.2
PING 10.53.1.2 (10.53.1.2) 56(84) bytes of data.
From 10.42.0.130 icmp_seq=1 Destination Host Unreachable
From 10.42.0.130 icmp_seq=5 Destination Host Unreachable
From 10.42.0.130 icmp_seq=6 Destination Host Unreachable
^C
--- 10.53.1.2 ping statistics ---
9 packets transmitted, 0 received, +3 errors, 100% packet loss, time 8191ms
pipe 4

At the core computer:

marco@marco-NUC7i7BNH:~$ sudo tcpdump -i eno1 icmp
tcpdump: verbose output suppressed, use -v[v]... for full protocol decode
listening on eno1, link-type EN10MB (Ethernet), snapshot length 262144 bytes
^C
0 packets captured
0 packets received by filter

Edit: I guess that the problem was related to the routes. I took out some of them and then I was able to ping 10.53.1.2 using the gNB computer

marco@darlene-G5-KC:~/Downloads$ ping 10.53.1.2
PING 10.53.1.2 (10.53.1.2) 56(84) bytes of data.
64 bytes from 10.53.1.2: icmp_seq=1 ttl=63 time=0.564 ms
64 bytes from 10.53.1.2: icmp_seq=2 ttl=63 time=0.508 ms
64 bytes from 10.53.1.2: icmp_seq=3 ttl=63 time=0.437 ms
^C
--- 10.53.1.2 ping statistics ---
3 packets transmitted, 3 received, 0% packet loss, time 2047ms
rtt min/avg/max/mdev = 0.437/0.503/0.564/0.051 ms
marco@darlene-G5-KC:~/Downloads$ sudo ip route
default via 10.42.0.1 dev enp8s0f1 proto dhcp metric 20100 
10.42.0.0/24 dev enp8s0f1 proto kernel scope link src 10.42.0.130 metric 100 
169.254.0.0/16 dev enp8s0f1 scope link metric 1000 
172.17.0.0/16 dev docker0 proto kernel scope link src 172.17.0.1 linkdown
pgawlowicz commented 4 months ago

Hi, could you try the following:

  1. Expose Open5gs ports in docker-compose.yml

     # Uncomment port to use the 5gc from outside the docker network
-    #   - "38412:38412/sctp"
+      - "38412:38412/sctp"
+      - "2152:2152/udp"

  2. Add advertise parameter in upf.gtpu in open5g config file. 

    gtpu:
 server:
   - address: ${OPEN5GS_IP}
+        advertise: x.x.x.x # IP address of the Open5gs container host PC

  3. Rebuild Open5gs container:

    docker compose up --build 5gc

  4. gnb config: 

    amf:
addr: x.x.x.x                 # IP address of the Open5gs container host PC
bind_addr: y.y.y.y         # IP address of the gnb host PC
mas-oliveira commented 4 months ago

I added these configurations but I'm still having an issue. Somehow it looks like that I receive the ICMP packets on the core computer:

marco@marco-NUC7i7BNH:~/srsRAN_Project/docker$ sudo tcpdump -i any icmp
[sudo] password for marco: 
tcpdump: data link type LINUX_SLL2
tcpdump: verbose output suppressed, use -v[v]... for full protocol decode
listening on any, link-type LINUX_SLL2 (Linux cooked v2), snapshot length 262144 bytes
14:45:14.203940 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
14:45:14.203942 vethef0918c Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
14:45:15.168505 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
14:45:15.168507 vethef0918c Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
14:45:16.205648 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
14:45:16.205651 vethef0918c Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
14:45:17.243759 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
14:45:17.243761 vethef0918c Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136

But I guess that there is some error with some routes. Which is weird since that from the same computer (marco-NUC7i7BNH) with the core I can ping 10.53.1.2:

marco@marco-NUC7i7BNH:~/srsRAN_Project/docker$ ping 10.53.1.2
PING 10.53.1.2 (10.53.1.2) 56(84) bytes of data.
64 bytes from 10.53.1.2: icmp_seq=1 ttl=64 time=0.056 ms
64 bytes from 10.53.1.2: icmp_seq=2 ttl=64 time=0.049 ms
^C
--- 10.53.1.2 ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1021ms
rtt min/avg/max/mdev = 0.049/0.052/0.056/0.003 ms

I'm still doing the same command on the UE side:

sudo ip netns exec ue1 ping 10.45.1.1
pgawlowicz commented 4 months ago

did you enable forwarding on both gnb and open5gs hosts?

sudo sysctl -w net.ipv4.ip_forward=1
sudo sysctl -w net.ipv4.conf.all.forwarding=1
sudo sysctl -w net.ipv6.conf.all.forwarding=1
sudo iptables -t nat -A POSTROUTING -o $OUT_INTERFACE -j MASQUERADE
sudo iptables -P FORWARD ACCEPT
mas-oliveira commented 4 months ago

Yes. When I ping via UE using:

sudo ip netns exec ue1 ping 10.45.1.1

I can't reach the core, and the packet capture on the core machine stills:

marco@marco-NUC7i7BNH:~/srsRAN_Project/docker$ sudo tcpdump -i any icmp
tcpdump: data link type LINUX_SLL2
tcpdump: verbose output suppressed, use -v[v]... for full protocol decode
listening on any, link-type LINUX_SLL2 (Linux cooked v2), snapshot length 262144 bytes
15:14:50.810214 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:14:50.810216 vethd20576b Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:14:51.849485 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:14:51.849487 vethd20576b Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:14:52.889749 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:14:52.889751 vethd20576b Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:14:53.849661 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:14:53.849664 vethd20576b Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:14:55.433121 lo    In  IP localhost > localhost: ICMP localhost udp port 48504 unreachable, length 194

But when I try to reach the core (10.53.1.2) on the machine that has the gNB but using just:

ping 10.53.1.2

I can reach the core:

marco@marco-NUC7i7BNH:~/srsRAN_Project/docker$ sudo tcpdump -i any icmp
tcpdump: data link type LINUX_SLL2
tcpdump: verbose output suppressed, use -v[v]... for full protocol decode
listening on any, link-type LINUX_SLL2 (Linux cooked v2), snapshot length 262144 bytes
15:15:10.119870 eno1  In  IP 10.42.0.130 > 10.53.1.2: ICMP echo request, id 4, seq 1, length 64
15:15:10.119912 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP echo request, id 4, seq 1, length 64
15:15:10.119916 vethd20576b Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP echo request, id 4, seq 1, length 64
15:15:10.119941 vethd20576b P   IP 10.53.1.2 > marco-NUC7i7BNH: ICMP echo reply, id 4, seq 1, length 64
15:15:10.119941 br-26141c6f72a5 In  IP 10.53.1.2 > 10.42.0.130: ICMP echo reply, id 4, seq 1, length 64
15:15:10.119950 eno1  Out IP 10.53.1.2 > 10.42.0.130: ICMP echo reply, id 4, seq 1, length 64
15:15:11.151961 eno1  In  IP 10.42.0.130 > 10.53.1.2: ICMP echo request, id 4, seq 2, length 64
15:15:11.151982 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP echo request, id 4, seq 2, length 64
15:15:11.151985 vethd20576b Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP echo request, id 4, seq 2, length 64
15:15:11.152008 vethd20576b P   IP 10.53.1.2 > marco-NUC7i7BNH: ICMP echo reply, id 4, seq 2, length 64
15:15:11.152008 br-26141c6f72a5 In  IP 10.53.1.2 > 10.42.0.130: ICMP echo reply, id 4, seq 2, length 64
15:15:11.152017 eno1  Out IP 10.53.1.2 > 10.42.0.130: ICMP echo reply, id 4, seq 2, length 64
pgawlowicz commented 4 months ago

Could you reset the PCs and start again with clean routing tables?

mas-oliveira commented 4 months ago

It stills happening.

15:52:17.166359 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:52:17.166361 veth9a7e1be Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:52:18.119984 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:52:18.119986 veth9a7e1be Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:52:19.161426 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:52:19.161429 veth9a7e1be Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:52:20.216601 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:52:20.216603 veth9a7e1be Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:52:21.238973 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
15:52:21.238974 veth9a7e1be Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136

While trying to reach 10.45.1.1 using UE1.

I added the set of rules to both computers:

sudo sysctl -w net.ipv4.ip_forward=1
sudo sysctl -w net.ipv4.conf.all.forwarding=1
sudo sysctl -w net.ipv6.conf.all.forwarding=1
sudo iptables -t nat -A POSTROUTING -o $OUT_INTERFACE -j MASQUERADE
sudo iptables -P FORWARD ACCEPT

Just to verify, the OUT_INTERFACE should be the interface to the ethernet cable that is being used to connect the computers, right?

I will show you my routing tables to each computer: gNB + UEs:

marco@darlene-G5-KC:~/srsRAN_4G/build/srsue/src$ route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
0.0.0.0         10.42.0.1       0.0.0.0         UG    20100  0        0 enp8s0f1
10.42.0.0       0.0.0.0         255.255.255.0   U     100    0        0 enp8s0f1
169.254.0.0     0.0.0.0         255.255.0.0     U     1000   0        0 enp8s0f1
172.17.0.0      0.0.0.0         255.255.0.0     U     0      0        0 docker0

Core:

marco@marco-NUC7i7BNH:~/srsRAN_Project/docker$ route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
0.0.0.0         10.3.10.1       0.0.0.0         UG    600    0        0 wlp58s0
10.0.2.0        0.0.0.0         255.255.255.0   U     0      0        0 br-0ae98bbb2186
10.3.10.0       0.0.0.0         255.255.255.0   U     600    0        0 wlp58s0
10.42.0.0       0.0.0.0         255.255.255.0   U     100    0        0 eno1
10.53.1.0       0.0.0.0         255.255.255.0   U     0      0        0 br-26141c6f72a5
169.254.0.0     0.0.0.0         255.255.0.0     U     1000   0        0 eno1
172.17.0.0      0.0.0.0         255.255.0.0     U     0      0        0 docker0
172.19.1.0      0.0.0.0         255.255.255.0   U     0      0        0 br-60b9fad393f0
pgawlowicz commented 4 months ago

could you remove the open5gs container and build it again? could you share your updated config files?

mas-oliveira commented 4 months ago

I did it using docker rm open5gs_5gc. I have rebuilt and then I had the same result.

Conf files:

docker-compose on srsRAN_Project docker-compose.zip

Open5GS yml, env and dockerfile open5gs_conf.zip

/etc/open5gs config files etc_5gs_confs.zip

gNB conf file gnb_zmq.zip

UE conf file ue1_zmq.txt

pgawlowicz commented 4 months ago

could you try:

    gtpu:
      - addr: ${OPEN5GS_IP}
        advertise: 10.42.0.1

instead of:

    gtpu:
      - addr: ${OPEN5GS_IP}
      - advertise: 10.42.0.1

and rebuild the container?

yc541 commented 4 months ago

I have followed the comments from @pgawlowicz and added a routing rule on the gNB machine _sudo ip route add 10.53.1.0/24 via core_machineIP Now my gNB machine can ping 10.53.1.2 and once the UE attaches it can ping the core and has Internet.

mas-oliveira commented 4 months ago

I changed the gtpu to:

      - addr: ${OPEN5GS_IP}
        advertise: 10.42.0.1

I can ping the core using the gNB computer:

marco@darlene-G5-KC:~/srsRAN_4G/build/srsue/src$ ping 10.53.1.2
PING 10.53.1.2 (10.53.1.2) 56(84) bytes of data.
64 bytes from 10.53.1.2: icmp_seq=1 ttl=63 time=0.686 ms
64 bytes from 10.53.1.2: icmp_seq=2 ttl=63 time=0.521 ms
64 bytes from 10.53.1.2: icmp_seq=3 ttl=63 time=0.636 ms
64 bytes from 10.53.1.2: icmp_seq=4 ttl=63 time=0.602 ms

But I can't reach the core using the UE anyway:

marco@darlene-G5-KC:~/srsRAN_4G/build/srsue/src$ sudo ip netns exec ue1 ping 10.45.1.1
PING 10.45.1.1 (10.45.1.1) 56(84) bytes of data.
^C
--- 10.45.1.1 ping statistics ---
58 packets transmitted, 0 received, 100% packet loss, time 58351ms

I can see on the core computer that when I ping 10.45.1.1 using the UE1 on tcpdump the traffic can't reach the host:

21:59:08.614329 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
21:59:08.614331 vethd209c53 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
21:59:09.618213 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
21:59:09.618216 vethd209c53 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
21:59:10.642752 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136
21:59:10.642755 vethd209c53 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP 10.42.0.130 udp port 2152 unreachable, length 136

Where I can confirm that on 10.42.0.130 (gNB machine) I have the gNB running on port 2152:

marco@darlene-G5-KC:~/srsRAN_4G/build/srsue/src$ sudo lsof -i -n | grep 2152
gnb       9001            root    8u  IPv4 121627      0t0  UDP 10.42.0.130:2152

So I don't know why it's unreachable... I don't have any firewall active by the way

While, when I ping directly 10.53.1.2 from the gNB computer everything goes well, and it seems that both are using the same interfaces to reach the core:

21:59:21.096312 eno1  In  IP 10.42.0.130 > 10.53.1.2: ICMP echo request, id 9, seq 1, length 64
21:59:21.096343 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP echo request, id 9, seq 1, length 64
21:59:21.096347 vethd209c53 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP echo request, id 9, seq 1, length 64
21:59:21.096374 vethd209c53 P   IP 10.53.1.2 > marco-NUC7i7BNH: ICMP echo reply, id 9, seq 1, length 64
21:59:21.096374 br-26141c6f72a5 In  IP 10.53.1.2 > 10.42.0.130: ICMP echo reply, id 9, seq 1, length 64
21:59:21.096384 eno1  Out IP 10.53.1.2 > 10.42.0.130: ICMP echo reply, id 9, seq 1, length 64
21:59:22.107648 eno1  In  IP 10.42.0.130 > 10.53.1.2: ICMP echo request, id 9, seq 2, length 64
21:59:22.107668 br-26141c6f72a5 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP echo request, id 9, seq 2, length 64
21:59:22.107672 vethd209c53 Out IP marco-NUC7i7BNH > 10.53.1.2: ICMP echo request, id 9, seq 2, length 64
21:59:22.107696 vethd209c53 P   IP 10.53.1.2 > marco-NUC7i7BNH: ICMP echo reply, id 9, seq 2, length 64
21:59:22.107696 br-26141c6f72a5 In  IP 10.53.1.2 > 10.42.0.130: ICMP echo reply, id 9, seq 2, length 64
21:59:22.107706 eno1  Out IP 10.53.1.2 > 10.42.0.130: ICMP echo reply, id 9, seq 2, length 64
pgawlowicz commented 4 months ago

@mas-oliveira could you use the newest docker-compose.yml, where we use Open5gs v2.7.0 instead of v2.6.1 ?

btw. why do you have /etc/open5gs config file? Do you run both manually installed open5gs and the dockerized version?

pgawlowicz commented 4 months ago

@mas-oliveira any update on this issue?

mas-oliveira commented 4 months ago

@mas-oliveira any update on this issue?

Sorry Piotr, I only will be able to give any update since tomorrow