Dynamic CTF changes trace levels without making broad tracing permanent
The Consolidated Tracing Facility (CTF) can collect diagnostic events from Rocket components and application-supplied CTF events. Its configuration determines which components are traced, the minimum severity captured, and where output is emitted. Text and binary emitters are supported; binary output can be examined with the CTF Viewer. Consolidated Tracing Facility (Diagnostics) [1]
Read the full briefing
Scope: Rocket Enterprise Developer 12.0 for Eclipse on Linux/UNIX; Enterprise Server diagnostics.
The Consolidated Tracing Facility (CTF) can collect diagnostic events from Rocket components and application-supplied CTF events. Its configuration determines which components are traced, the minimum severity captured, and where output is emitted. Text and binary emitters are supported; binary output can be examined with the CTF Viewer. Consolidated Tracing Facility (Diagnostics)
The normal, initialization-time configuration uses a text file—commonly ctf.cfg—identified by MFTRACE_CONFIG. For an Enterprise Server region, set this environment variable under [ES-Environment] in the region’s Configuration Information. The file is processed during run-time initialization and contains one name-value pair per line; lines beginning with # are comments. Configuration File for CTF
Tracing can be narrowed by component:
mftrace.level.mf.rts=debugDEBUG captures every event and is the most verbose threshold. Other documented levels are INFO, WARN, ERROR, FATAL, and NONE. Component names and level values are case-insensitive. Levels also inherit hierarchically: if mf.rts.subsystem has no explicit level, it can inherit from mf.rts, which in turn can inherit from mf. A later entry for the same component overrides an earlier one. mftrace.level.component-name
For a running region, ESCWA’s ES Control page provides a separate dynamic mechanism. Set Dynamic CTF File Location, then choose a CTF Action:
- As on file applies the levels specified in that dynamic configuration.
- All on sets
DEBUGfor every component having anmftrace.level.component_nameentry in the file. - All off disables tracing for those listed components.
Output continues to the location established by the original startup configuration when MFTRACE_CONFIG was set before region startup; otherwise, it goes to the region’s working directory. This lets an operator prepare a narrowly scoped component list and activate detailed tracing only while reproducing a problem. ES Control
The critical safety rule is explicit: set CTF Action to All off before changing either the dynamic file location or the file’s contents. Rocket warns that failing to do so results in memory corruption. Also, All on forces every listed component to DEBUG, regardless of the finer levels written in the file; use As on file when those thresholds matter.
Takeaway: Treat the dynamic CTF file as an activation manifest—disable it before editing, list only the required components, and use As on file when you need controlled severity rather than blanket DEBUG.