🔧 Field Note Oct 18, 2026

Hands-On Scenario: Attach OK, No Data

🎯 Scenario

A subscriber attaches successfully and gets an IP address. Signal is good. But no application works: no browsing, no messaging, no ping replies. The radio team says “radio is fine”. It is your case.

📋 Symptoms (given)

Attach Accept received, default bearer is up.
UE has an IP address from the PGW.
Downlink and uplink packet counters on the UPF are near zero.
No attach rejects, no alarms on the MME.
Other subscribers on the same eNodeB work fine.

🔍 Investigation Steps

Work in order. Do not skip ahead.

Step 1 — Confirm the bearer is really up

Check the default bearer in the MME and SGW. Is it Active, or stuck in a transitional state? Which APN did the UE actually get?

Step 2 — Check DNS

From the UE’s perspective: does DNS resolve? Query the APN’s DNS records (NAPTR/SRV) from the MME side. A failed, stale, or wrong APN DNS record is the classic “full bars, zero data” fault.

Step 3 — Check the PGW and charging

Is the session alive in the PGW? Did the OCS answer the Gy CCR, or is the session blocked on credit (Result-Code 4012)? Check the Gy exchange and the FUI action if quota is exhausted.

Step 4 — Check the user plane path

In the UPF (or SGW-U/PGW-U): are the PDR/FAR rules installed for this session? Is the F-TEID in the PDR correct? A missing or mismatched F-TEID drops packets silently.

Step 5 — Check policy

Did the PCRF/PCF return a policy that gates or blocks this APN or subscriber? Compare the policy the SMF/PGW received with what is enforced.

❓ Questions

At which step did the chain break, and what evidence proves it?
If DNS is the fault, what exactly is wrong: missing record, stale record, or wrong priority?
If Gy is the fault, what did the OCS answer, and what should the PGW do with the bearer?
If PFCP is the fault, which rule is missing or mismatched?

✅ Answer Key

Most likely fault: DNS APN resolution (Step 2) or Gy credit block (Step 3). Both produce “attach OK, no data” with a clean radio picture.
DNS evidence: the MME’s APN lookup fails, returns a stale record, or points at a decommissioned PGW.
Gy evidence: CCA with Result-Code 4012, or FUI = Terminate / Redirect. If FUI is Redirect, the UE should reach the top-up server and nothing else.
PFCP evidence: session exists in the SMF but the UPF has no matching PDR/FAR, or the F-TEID is wrong. Fix: re-establish or modify the PFCP session with correct rules.
Rule: find the first step that broke. Everything after it is noise.

🧠 PacketAlchemy Note

“Attach OK, no data” is not one fault. It is a symptom with at least four different causes. The skill is not knowing the answer — it is working the chain in order until the first break appears.

Lab Troubleshooting EPC DNS Gy PFCP