ci-server-engineer/.claude/skills/system-management/tasks/bus-utilities.md

task	version	source	status
BUS Service utilities - process checks, stopping single processes, message tools, adaptive queues	R1.05 (first written for R1.03)	IM 36K10L30-01EN §1, §5, §6, §10, §14 (4th ed., Feb. 2026); IM 36K10J10-01EN §3.2.6 (5th ed.); checked on the test system 2026-10-01 (R1.03)	verified on R1.03 (durwfp, setup file checks, umhman, umhmsg, umhtestprog); re-checked read-only on the R1.05 test server 2026-10-08 (durwfp exit codes, umhman, dur.sup, durm_1_udp.sup); umhmsg / umhtestprog not run on R1.05 (they write to the shared message log); durstp not run

BUS Service utilities

Background: knowledge/manuals/IM36K10L30-01EN.md.

Is a CI Server process attached?

durwfp -p <DUR name> -t 0 → exit code 0 = attached (running), 1 = not. Without -t 0 it waits up to 60 s for the process to appear; -e waits for it to go away. Verified: umh, opcman, durm_1 → 0; unknown name → 1 (R1.03 and R1.05; on R1.05 also dssman, dsssda00, dsssdarst, dsssdapy, dssevt00 → 0). Bash: durwfp -p opcman -t 0; echo $?; PowerShell: durwfp -p opcman -t 0; $LASTEXITCODE. The manual (L30 §10, reworded in the R1.04 edition) calls the result "1 if attached, else 0": that is the opposite of the exit code a script sees, so test the exit code as above (ISS-120). DUR names are those given with -n (e.g. OPXCIENG, EQPCIENG, opxuas).

Stop or kill one process (only with the user's approval)

durstp -m -p <name> sends a stop message (clean stop); durstp -m -k -t 10 -p <name> stops, then kills after 10 s. * wildcards stop many processes (-p alm\*). Never durstp -k -p \* (kills durtim and leaves DUR memory locked). Restart a process by hand with <exe> -n <name> from tls\exe or by restarting CI Server.

Messages

• Message log: %TLS_LOG_PATH%\umhlog.log (--> From: <PROCESS> (<node>) on <time>, then FAC-S-CODE, text lines). UMH Message Text Error: Message number <hex> (…) = the text could not be formatted (e.g. % in a PRINT text, ISS-025).
• umhman (hex code → text): answers can be piped, e.g. printf '8010008\n\n' | umhman → DUR-W-NO_MSG_FND, No selected message found; parameters in the text stay empty. Same answer on R1.05 (piped from PowerShell, 2026-10-08).
• umhmsg <hex> [<var> <value>]… logs that message from a script: it prints nothing; the message appears in umhlog.log from process UMHMSG_<pid>.
• umhtestprog runs its self-test (TESTING UMH MESSAGES, TESTING ILLEGAL CODES).
• On an unlicensed node these tools crash or stop with license expired once the licence time is nearly used up, while dssqld/itmval still work; restart CI Server (ISS-030).

Adaptive queues (read-only check)

tls\sup\dur.sup, tab Adaptive queues: ADAPTIVE_QUEUE_SETS = <process>, <percent> (wildcard * allowed, max 40 %). Installed: ALM 10, ALH 10, WFLGDI 20, WFLCDI, WFLIDI, WFLADI, ITHEIX, ITHSER 2, DSSSDA*, DSSEVT*, OPCEXE 1; R1.05 adds ADTSTO 1 (audit storage, added "for version upgrade" per the file's change log; ISS-122). Add entries for new busy processes (extra executors, equipment managers, custom processes); OPCEXE without * may not cover the installed OPCEXE1…4 (ISS-026; unchanged on R1.05). Changing dur.sup needs admin and a restart (skill portal-setup). Queue use per process: health snapshots (tuning-checks.md) or the performance monitor (Minimum/Maximum/Current KB, high queue 100 % = the queue had to grow → raise its size).

Gateway Core switch (R1.05)

R1.05 dur.sup has a new keyword GATEWAY_CORE (General tab, Gateway Core check box; installed gateway_core = No). Tick it only on a CI Core with a gateway licence: such a node collects and passes on data but has no operator interface, also not for Portals and CI View clients connected to it. If the switch and the licence disagree, a message goes to the message log (IM 36K10J10 §3.2.6). From docs and the installed file; not tested (no gateway licence; changing dur.sup needs admin and a restart; ISS-116).

Node links (MDUR)

tls\sup\durm_<line>_udp.sup and tls\com\durm_start.cmd; single-node installs still run durm_1 (durm_1_udp.sup: line 1, quality 500, port 17001). Multi-node set-up is in the L30 summary §2; not practiced (one node).