Closed josebeo2016 closed 1 year ago
Hi @josebeo2016 , I'm having a similar issue. Have you managed to find the fix for this? Thank you and brs, Tamas
Hi @josebeo2016 , I'm having a similar issue. Have you managed to find the fix for this? Thank you and brs, Tamas
I installed both Calico and Cilium working together in my system and the above problem is over. However, I face new issue:
$ >> k -n free5gc logs free5gc-v1-free5gc-upf-upf-658c9d59-qprgg
Cannot find device "n6"
2022-03-08T02:25:45Z [INFO][UPF][Util] Config: /free5gc/config/..2022_03_08_02_25_01.1342527809/upfcfg.yaml
2022-03-08T02:25:45Z [INFO][UPF][Util] UPF config version [1.0.0]
2022-03-08T02:25:45Z [INFO][UPF][Util] Set log level: info
2022-03-08T02:25:45Z [ERRO][UPF][Util] Socket bind fail : Cannot assign requested address
2022-03-08T02:25:45Z [ERRO][UPF][Util] Socket -1 register event in epoll error : Bad file descriptor
2022-03-08T02:25:45Z [ERRO][UPF][Util] UPDK epoll register error
2022-03-08T02:25:46Z [ERRO][UPF][Util] Pool is full, it may not belong to this pool
2022-03-08T02:25:46Z [ERRO][UPF][Util] Gtp5gDeviceAdd failed
2022-03-08T02:25:46Z [ERRO][UPF][Util] Socket bind fail : Cannot assign requested address
2022-03-08T02:25:46Z [ERRO][UPF][Util] Socket -1 register event in epoll error : Bad file descriptor
2022-03-08T02:25:46Z [ERRO][UPF][Util] PFCP Sock Register to epoll error
2022-03-08T02:25:46Z [ERRO][UPF][Util] Create PFCP Server for IPv4 error
2022-03-08T02:25:46Z [ERRO][UPF][Util] UPF - PFCP error when UPF initializes
2022-03-08T02:25:46Z [ERRO][UPF][Util] UPF failed to initialize
2022-03-08T02:25:46Z [ERRO][UPF][Util] Pool is full, it may not belong to this pool
2022-03-08T02:25:46Z [INFO][UPF][Util] Removing DNN routes
2022-03-08T02:25:46Z [ERRO][UPF][Util] if_nametoindex
2022-03-08T02:25:46Z [ERRO][UPF][Util] Delete routing rule to device upfgtp failed: 10.1.0.0/17
2022-03-08T02:25:46Z [ERRO][UPF][Util] UPDK epoll deregister error
I'm working now to fix the GTP5g module in my worker node now. I will update soon
I'm sorry Cilium has never been tested as the only and primary CNI plugin on the cluster. For the GTP5g kernel module, I'm using the tag v0.3.2 of this repo https://github.com/free5gc/gtp5g.git and it works well
Thank you for your suggestion. I will try to install the gtp5g v0.3.2. Last work, I tried the latest gtp5g version with self-modification on the original gtp5g source code because I use the Linux kernel 5.11.0.
Okay let me know about the result of gtp5g v0.3.2
Hi
Did this work for you? I am closing the issue but don't hesitate to re-open it if necessary
Raouf
It is not because of gtp5g. Network Interface in Ubuntu 18.04 is not reprented by eth0, so some configuration about network in grub
are required changing.
Here is a temporary solution.
ip link set ens37 name eth0
ip link set eth0 up
ip link set ens38 name eth1
ip link set eth1 up
ens37 and ens38 are network interfaces in Ubuntu 18.04.
Another solution by changing file /etc/default/grub
GRUB_CMDLINE_LINUX="net.ifnames=0 biosdevname=0"
and then execute sudo grub-mkconfig -o /boot/grub/grub.cfg
The reason is that eth0 and eth1 are used to configure N3, N4, N6 etc.
Hello everyone.
I'm trying to install towards5gs-helm on my system. The specs: Kubernetes
Network configuration Because the names of network interfaces on your Kubernetes nodes are different from eth0 and eth1, I changed all these parameter into the name of my network interface by modifying the
towards5gs-helm/charts/free5gc/values.yaml
:I also tried to change other similar config of the amf, smf and upf chart. But the problem still happens.
The error
kubectl get pods -n free5gc
kubectl describe pod -n free5gc free5gc-v1-free5gc-amf-amf-89966854-xq5r4