docs(index): OtOpcUa c9640f5c deployed to DEV vd03 + QA vq03/vq04

Claude-Session: https://claude.ai/code/session_01GT67ACmpMR7b9LGRPLijMi
This commit is contained in:
Joseph Doherty
2026-11-13 02:58:23 -05:00
parent 127f07e1ef
commit 65197ca1a2
2 changed files with 24 additions and 2 deletions
@@ -584,3 +584,25 @@ vq03 **250** / vq04 **240**. Record: OtOpcUa `docs/deploy/qa-upgrade-b8b4c18f-to
sick runner shows as `BadResourceUnavailable` on those two nodes (ScadaBridge#112). ScadaBridge#125: the 2.0 deferral
condition (#799 + #801 closed) is now met.
### 2026-11-13 — DEV vd03 (`6c4d2990`) and the QA pair (`fea7de0a`) upgraded to master `c9640f5c` (#901, #902, #903, #834, #872, #904)
Owner-approved; WinRM only for the pack's on-box steps. Pack `docs/deploy/2026-11-13-next-vd03-qa-to-c9640f5c.md`; record
`docs/deploy/2026-11-13-vd03-qa-to-c9640f5c.md` (`c0ba450d`). One self-contained build (1398 files, zip `12C46E60…`, Host.dll `aff960bd2d36`).
- **#901 first real use.** On each node, one full-argument installer run wrote the settings record (`SettingsRecord` changed; on QA also a
reorder-only `HostServiceEnv`, sorted env SHA unchanged). DenyPaths were kept exactly as recorded (vd03 5, QA 12 incl. OtelCollector).
Swaps then ran through `Invoke-OtOpcUaInstallSwap.ps1`. `-Reapply` restored exactly the ACEs the move drops; `-Verify` OK (13/13 vd03, 21/21 QA)
after every swap and every start.
- **Downtime is longer than estimated:** vd03 ~101 s, QA pair ~91 s. The coverage walk inside `-Reapply`/`-Verify` runs while the service is
stopped (on vd03 the deny covers all of `E:\ApiInstall`). On QA the Secondary was swapped (`-NoStart`) while the Primary still served, which
kept one swap out of the pair outage.
- **Gates:** vd03 SL 250, FOCAS 21/22 (BlockCount BadNoData), #834 preflight line "every stored and deployed reference is namespaced", 12
aggregates, HG Probe only. #902 PASS: no `[ERR]` on a Monaco tab close; the old build logged one. #903 PASS: "Checking for changes…" → drift
without interaction. QA 250/240; failover vq03→vq04 at 250 on the first sample; vq03 back at 240, no failback; LocalDb connected with backlog 0;
canary Good on both; `worker_up` 1 per shard; HG Probe only.
- **Finding (#903 comment):** the revision hash includes each row's `RowVersion`, so an edit-and-revert leaves the drift chip on
"Configuration drift" until the next deploy. vd03 is left that way: content is restored, the running config is still `fa9a2e28…`, and no deploy was made.
- QA Primary is **vq04** after the run. Rollback: vd03 `E:\ApiInstall\_stage-otopcua-c9640f5c\rollback.ps1` (prev `OtOpcUa.bak.20261113-pre-c9640f5c`);
QA `E:\psrem-stage\otopcua-c9640f5c\rollback.ps1` per node (prev `app\..\app.bak.20261113-pre-c9640f5c`). DB `.bak`s on both SQL servers.
- **ScadaBridge consumer impact: none on DEV/QA** (no OPC UA link). Elsewhere: see the pack §1. Galaxy Pending/Warning now reads Uncertain
(#872), and an out-of-namespace credential ref no longer connects (#834).
+2 -2
View File
File diff suppressed because one or more lines are too long