With latest compile srsRAN 5G package, getting AMF connection fail

[wei-lou]

I installed latest srsRAN 5g package via local compiling from source code. While I execute the gnb command, i am not able to get it connect to AMF. However the older version of srsRAN 5g (srsRAN gNB (commit 0112729)) works fine.

pc@srsRAN5G:~$ sudo gnb -c srsRAN5G/configs/gnb_ru_fujin78_tdd_n78_100mhz_2x2.yml
Could not set the affinity for the ru_rx_0 worker

--== srsRAN gNB (commit 5e6f50a) ==--

Connecting to AMF on 192.168.70.210:38412
Failed to connect SCTP socket to 192.168.70.210:38412. error="Network is unreachable" timeout=0ms
srsGNB ERROR: Failed to create SCTP gateway.

Aborted (core dumped)

tail: /tmp/gnb.log: file truncated
2023-10-10T15:53:59.106417 [GNB ] [I] Built in Release mode using commit 5e6f50a on branch main
2023-10-10T15:53:59.106434 [GNB ] [D] Required CPU features: sse4.1(ok) avx(ok) avx2(ok) avx512f(ok) avx512bw(ok) avx512vl(ok) avx512dq(ok) avx512cd(ok) fma(ok) pclmul(ok)
2023-10-10T15:53:59.106484 [GNB ] [W] Could not check scaling governor. filename=/sys/devices/system/cpu/cpu0/cpufreq/scaling_governor error="No such file or directory"
2023-10-10T15:53:59.106519 [GNB ] [D] DRM KMS polling is disabled
2023-10-10T15:53:59.107542 [ALL ] [I] Task worker "NGAP-PCAP" started...
2023-10-10T15:53:59.107687 [ALL ] [I] Task worker "E1AP-PCAP" started...
2023-10-10T15:53:59.107714 [ALL ] [I] Task worker "GTPU-PCAP" started...
2023-10-10T15:53:59.107827 [ALL ] [I] Task worker "F1AP-PCAP" started...
2023-10-10T15:53:59.107837 [ALL ] [I] Task worker "E2AP-PCAP" started...
2023-10-10T15:53:59.108153 [ALL ] [I] Task worker "MAC-PCAP" started...
2023-10-10T15:53:59.108983 [ALL ] [I] Task worker "ru_timing" started...
2023-10-10T15:53:59.109179 [ALL ] [I] Task worker "ru_dl_#0#0" started...
2023-10-10T15:53:59.109318 [ALL ] [I] Task worker "ru_tx_0" started...
2023-10-10T15:53:59.110127 [ALL ] [I] Task worker "ru_rx_0" started...
2023-10-10T15:53:59.110251 [ALL ] [I] Task worker "gnb_ue" started...
2023-10-10T15:53:59.111765 [ALL ] [I] Task worker "phy_prach#0" started...
2023-10-10T15:53:59.111931 [ALL ] [I] Task worker "up_phy_dl#0#0" started...
2023-10-10T15:53:59.112828 [GNB ] [I] Connecting to AMF (192.168.70.210)..
2023-10-10T15:53:59.113012 [SCTP-GW ] [D] Setting RTO_INFO options on SCTP socket. Association 0, Initial RTO 120, Minimum RTO 120, Maximum RTO 500
2023-10-10T15:53:59.113017 [SCTP-GW ] [D] Setting SCTP_INITMSG options on SCTP socket. Max attempts 3, Max init attempts timeout 500
2023-10-10T15:53:59.113046 [SCTP-GW ] [D] Binding to 127.0.0.1 port 0
2023-10-10T15:53:59.113065 [SCTP-GW ] [D] Binding successful
2023-10-10T15:53:59.113091 [SCTP-GW ] [D] Setting RTO_INFO options on SCTP socket. Association 0, Initial RTO 120, Minimum RTO 120, Maximum RTO 500
2023-10-10T15:53:59.113094 [SCTP-GW ] [D] Setting SCTP_INITMSG options on SCTP socket. Max attempts 3, Max init attempts timeout 500
2023-10-10T15:53:59.113098 [SCTP-GW ] [D] Connecting to 192.168.70.210 port 38412
2023-10-10T15:53:59.113175 [SCTP-GW ] [D] Failed to connect to 192.168.70.210:38412 - Network is unreachable
2023-10-10T15:53:59.113176 [SCTP-GW ] [D] Closing socket.
2023-10-10T15:53:59.113203 [SCTP-GW ] [E] Failed to connect SCTP socket to 192.168.70.210:38412. error="Network is unreachable" timeout=0ms

[brendan-mcauliffe]

Hi @wei-lou , are you sure the addr value you are using is correct and that the core is online? This error is usually only seen when there is a problem with the addr being used or the core itself.

[wei-lou]

Thanks for the response. Here are the logs.

CMakeCache.txt
gnb.log
gnb_ru_fujin78_tdd_n78_100mhz_2x2.zip

[ismagom]

Hi @wei-lou ,

Can you double-check your bind_addr? it doesn't seem to be correct. It should point to an IP that is accessible from the AMF. If the AMF is running in 192.168.70.0 subnet, it won't be able to access 127.0.0.1

  addr: 192.168.70.210                                            # The address or hostname of the AMF.
  bind_addr: 127.0.0.1                                            # A local IP that the gNB binds to for traffic from the AMF.

[wei-lou]

Hi @ismagom thanks for replying. I am confused here. I assume "addr" is the IP for the AMF (x.x.70.210). I am not sure bind_addr is referring to. Is this the IP for the GNB (in my case x.x.50.205 N2 interface)? The same configuration file which has bind_add as "127.0.0.1" with old version srsRAN has no issue.

[ismagom]

You need to set the IP of the GNB that can be reached from the AMF (try do pimg from AMF)

[wei-lou]

got it, thank u. Just wondering why the old version of software doesn't complain about the default localhost ip.

[arpitmal]

Do we have any update on this topic? I have used latest code and ending up in same issue. It works with old version, but do not work with latest code.

[ismagom]

Which is your error? Can you share your config file? Thanks

[arpitmal]

{
"amf": {
"addr": "172.24.232.68",
"bind_addr": "127.0.0.1"
},

eth0: flags=4419<UP,BROADCAST,RUNNING,PROMISC,MULTICAST> mtu 1500
inet 172.24.232.68 netmask 255.255.240.0 broadcast 172.24.239.255
inet6 fe80::215:5dff:feb4:efdf prefixlen 64 scopeid 0x20
ether 00:15:5d:b4:ef:df txqueuelen 1000 (Ethernet)
RX packets 177896 bytes 318412777 (318.4 MB)
RX errors 0 dropped 0 overruns 0 frame 0
TX packets 112523 bytes 133392235 (133.3 MB)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0

Both core and srsGnb running on same machine (but on different WSL versions gnb on 22.04, and core on 20.04) , but get below error, seems its not able to connect to fd created in create_and_bind seems.

2023-10-24T14:12:09.118498 [SCTP-GW ] [E] Failed to connect SCTP socket to 172.24.232.68:38412. error="Connection timed out" timeout=1374ms

[ismagom]

Look at my comment above. Check again the amf_bind address

[arpitmal]

ok got it, updated ip and its working fine.