ci-server-engineer/.claude/skills/history-management/tasks/history-checks.md
cliu 37c91eeac7 Add history manual summaries (L20, C40) and history-management skill
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-01 13:02:32 +10:00

46 lines
2.5 KiB
Markdown

---
task: Check storage groups, units and history clients with hisdbg; force a rollover
version: R1.03
source: IM 36K10L20-01EN §2.1, §3.1; IM 36K10C40-01EN §5.1; practiced on the test system and CI-Eng-Sample 2026-10-01
status: verified
---
# History checks (`hisdbg`)
Interactive; pipe the answers from bash with a timeout. The first answer is the node (blank =
local), the last `-1` exits. Never send blank lines at the action prompt blindly (a blank repeats
the previous action).
```
printf '\n1\n20\n6\n-1\n' | timeout 20 hisdbg
printf '\n21\n<GROUP>\n3\n\n-1\n' | timeout 20 hisdbg
```
| Option | Shows | Safe |
|---|---|---|
| 1 | Statistics: group count, logged-in clients, unit starts / stops / rollovers / deletions since start | yes |
| 20 | All groups: name, node,group number, history type (24 item, 21 alarm, 6 audit, 1058 report, 36 system log, 4 station statistics) | yes |
| 21 `<name>` | Group info: path, next unit and file name, opt_type (`1 1` Scan/Time, `2 2` Event/Item, `4 4` Aggregation), start, stop, first rollover, interval, life time (all GMT), GMT/LCT rollover, archiving | yes |
| 3 (then blank = this group) | Units: unit, data, file, start, stop or `--- active ---`, delete time | yes |
| 6 | Clients logged in at the history manager | yes |
| 11 | Unknown history files | yes |
| 5 | Force rollover of the current group (closes the active unit, starts the next) | changes files — with approval |
| 4, 8, 10, 12, 9 | Delete unit / group, restore unit, delete unknown files, change schedule | **no — never without explicit approval** |
History files are in `%TLS_ALLUSERS_PATH%\tls\his\` (path shown by option 21):
`h<node><group><unit><data>.his` with base-32 digits, e.g. group 1,54 → `h0101m…`.
## What to look for
- New group but no history: no active unit after loading → restart CI Server (ISS-006).
- Unit start dates: the first unit of an Event/Item or Aggregation group is dated 2 days back
(`gap_time`); history loaded before the first unit opened after the restart is dropped
(see `storage-groups.md` §5).
- After a CI Server restart the unit sequence increases (the start script runs `hisrep`).
- Delete times = stop time + life time.
## Verified (2026-10-01)
45 groups, 30 clients; `CIENG_EVENT` group 1,54, opt_type `2 2`, daily GMT rollover, life 7
days; units 1,1 (created at load) → 3,1 after restart → forced rollover closed 3,1 at 12:05:26
and opened 3,2. `CIENG_SCAN10S` 1,55 `1 1`; `CIENG_AGG` 1,56 `4 4` with extra `h0101ozzz…` files.