SMGW — Gateway Monitor for RFC Connections
SMGW is the Gateway Monitor, used to inspect and manage the RFC gateway process (gwrd) on an application server instance. It shows active connections, registered RFC server programs, the gateway log, and the enforcement status of the secinfo and reginfo access control files. It is the tool for diagnosing RFC-level connection failures that SM59 alone cannot explain.
This page covers SMGW, the gateway monitor used to troubleshoot RFC connections, registered programs, and gateway-level security enforcement. It focuses on distinguishing gateway-level blocks from RFC destination misconfiguration, and on where the gateway's data actually lives since most of it is not stored in database tables.
Reviewed by an ERPClimb SAP consultant on 15 Sept 2026· 1,148 words
What it does
SMGW opens the Gateway Monitor, the interface to the gwrd process that handles all RFC and CPI-C communication in and out of an application server instance. It lists currently logged on clients, active connections, registered external server programs, and the gateway's own trace log. It also shows whether the secinfo and reginfo access control lists are active, and in what enforcement mode. The one structural fact that explains most of the confusion around SMGW: almost nothing it displays is persisted in a database table. Connections, registrations and logs are live process state or OS-level files. Restarting the instance or the gateway process wipes the view clean, and two consultants looking at SMGW on two different application servers can see completely different pictures of the same logical system.
When it is used
SMGW gets used mid-incident, when an RFC call fails, a third-party or external program cannot register with the system, a batch job that spawns an external process hangs, or a printing/spool RFC connection drops. It sits below SM59: SM59 defines and functionally tests an RFC destination, but SMGW shows what actually happens at the transport layer once a connection attempt reaches the gateway. Reach for SMGW when SM59's connection test fails with a generic communication error, when an interface partner reports 'registration of tp ... failed' or 'no more connections', or when investigating whether the secinfo/reginfo access control files are blocking a legitimate external program from registering. There is no Fiori equivalent; this remains a classic SAP GUI Basis tool.
How to use it
- Run SMGW on the application server instance where the problem is actually occurring, not an arbitrary one, since gateway state is per instance.
- Use Goto to view Logged On Clients and Connections to see who or what is currently connected and for how long.
- Use Goto to view the Gateway Log to see recent connection and registration attempts, including denials.
- Check Goto Attributes or the security-related menu entries to see whether secinfo and reginfo are loaded and whether enforcement is active or log-only.
- Select a specific connection and cancel it if it is hung and blocking other registrations.
- Raise the gateway trace level temporarily if the log does not show enough detail to explain a denial, then reproduce the failing call.
Key fields
- No standard dictionary table stores the SMGW connection or registration list; it reflects live gwrd process memory and disappears on process restart.
- The gateway log is written to an OS-level trace file associated with the gwrd process, not to a database table.
- The secinfo and reginfo access control lists are OS-level text files read by the gateway at startup and on reload, not database entries.
- System log entries related to gateway denials can appear in the general system log stream, which SM21 reads from OS log files rather than a table.
How to prove it in the data
Because the gateway holds no database record of its own, proving a symptom means correlating two file-based sources rather than running a table query. Pull the gateway trace file for the instance and the time window of the failure and look for the offending program ID, host, or destination name and the denial reason text. Cross-check the same timestamp window in SM21's system log for a matching entry, since a gateway-level denial usually also raises a system log message. Separately, inspect the current content of the secinfo and reginfo files on that instance to confirm whether the calling host or program is actually listed as permitted.
ECC vs S/4HANA
The gateway process and SMGW itself are essentially unchanged between ECC and S/4HANA; the underlying gwrd architecture and the secinfo/reginfo security model carry forward without modification. There is no Fiori app that replaces SMGW, since gateway diagnosis remains a Basis-level, instance-specific task best done in SAP GUI. Consultants moving from ECC do not need to relearn anything structural here.
Common pitfalls
- Secinfo/reginfo enforcement gap: after a security hardening change or upgrade, the gateway starts enforcing access control files that were previously permissive, and a legitimate RFC server program is suddenly denied registration. Symptom is a 'registration of tp ... denied' entry in the gateway log. Check the enforcement mode first, then check whether the calling program's host and name are listed in the ACL files. The reflex fix of disabling gateway security entirely removes the protection for every other connection and should not be used as a shortcut.
- Connection or resource exhaustion: the gateway hits a connection limit and starts rejecting new registrations with a 'no more connections' style error. Check the current connection count against the configured maximum before restarting anything. Restarting the gateway process clears the symptom temporarily but does not fix a program that opens connections and never closes them, and the problem returns.
- Confusing SMGW with SM59: an RFC destination test in SM59 fails, and time gets spent reconfiguring the destination when the actual block happened at the gateway before the destination logic ever ran. If the gateway log shows a denial for the same connection attempt, the destination configuration is very likely fine.
- Wrong instance in view: SMGW is opened on an application server that never handled the failing call, because logon load balancing routed the real attempt elsewhere. Confirm which instance actually processed the connection before drawing conclusions from an empty or unrelated log.
- Trace level too low to see the reason: the gateway log shows that a connection was denied but not why. Raising the trace level and reproducing the failure is usually faster than guessing from the sparse default log.
Whose problem this is
This is Basis territory. A functional consultant may report the symptom, an ABAP developer may report a failing RFC-enabled function module call, but diagnosing and fixing gateway registration or ACL issues belongs to Basis. A good handover includes the exact gateway log lines with timestamps, the RFC destination or program name involved, and the calling host or IP.
Related SAP objects
Reviewed pages this object connects to in the ERPClimb knowledge graph.
Source: ERPClimb — https://erpclimb.com/sap-tcodes/smgwERPClimb is an independent platform and is not affiliated with SAP SE. Reference pages are written and reviewed by SAP consultants for learning and troubleshooting.