Hello Kamailio Community,

*This issue is being re-submitted with CDP debug logs for further analysis.*

We are experiencing a *critical issue* with Kamailio where the *SCSCF randomly stops sending CER messages to the HSS* in some cases, even after TCP connections are successfully established. This results in *complete authentication failures*.

Please find the (IMS containers + HSS) logs with and the diameter config through the following link:

https://www.filemail.com/d/vhdnkujjnernogn

*Environment*

- *Kamailio Version:* kamailio 6.0.2 (x86_64/linux) d4dc5d - *Deployment:* SCSCF and HSS as a container on same physical machine using macvlan networking

*Components:*

- SCSCF: 172.22.1.20 - HSS: 172.22.1.3 - ICSCF: 172.22.1.19

*Problem Description*

The issue occurs *randomly*. The system works normally for a period of time, then suddenly the Diameter connection between the SCSCF and HSS fails and enters what appears to be a *reconnection loop*.

*Here's the cycle we're observing:*

*1. *ERROR: ims_auth [cxdx_mar.c:132]: async_cdp_callback(): Transaction timeout - did not get MAA

*2. *SCSCF: WARNING: cdp [peermanager.c:337]: peer_timer(): Inactivity on peer [hss.epc.mnc092.mcc418.3gppnetwork.org] and no DWA, Closing peer...

*3. State Timeout (~40 seconds later):* S-CSCF: ERROR: cdp [peerstatemachine.c:164]: sm_process(): in state Closed timeout event Timeout

*4. Reconnection Attempt (~30 seconds later):*

SCSCF: INFO: cdp [peerstatemachine.c:551]: I_Snd_Conn_Req(): Peer hss.epc.mnc092.mcc418.3gppnetwork.org

SCSCF: INFO: cdp [receiver.c:991]: peer_connect(): Trying to connect to 172.22.1.3 port 3870

SCSCF: INFO: cdp [receiver.c:1090]: peer_connect(): Peer hss.epc.mnc092.mcc418.3gppnetwork.org:3870 connected

*5. HSS Side (waiting indefinitely):*HSS: [hss:0.19.37] 2025/12/25 04:46:39.060208 [INF] [scscf>hss:0][waiting for CER from 172.22.1.20:33284]

*Cascading Effects:*

*ICSCF accumulates branches:*

ICSCF: ERROR: tm [t_fwd.c:792]: add_uac(): maximum number of branches exceeded

ICSCF: ERROR: tm [t_fwd.c:1771]: t_forward_nonack(): failure to add branches

As a result, *UEs are unable to register to IMS*, leading to a *VoLTE service outage for all UEs*.

Thank you and I look forwarding to hearing from you