diff --git a/docs/session-ledger.md b/docs/session-ledger.md index 91b6b3a..78c4c27 100644 --- a/docs/session-ledger.md +++ b/docs/session-ledger.md @@ -706,9 +706,17 @@ **TWO CARRY-FORWARDS (deliverables CLOSED; these are NOT -- do not mark them done):** 1. **Boot-race proof.** `site-baseleg-office1.service` is installed + enabled, but the `After=libvirtd`+poll - race is UNPROVEN until an ACTUAL reboot self-heals `ssh voffice1` with no manual step. The next natural - reboot is the test; if `ssh voffice1` is dead after any reboot, `bash scripts/site-baseleg.sh check - office1` is the one-command diagnosis (skill routing row). + race is UNPROVEN until an ACTUAL reboot self-heals `ssh voffice1` with no manual step. If `ssh voffice1` + is dead after any reboot, `bash scripts/site-baseleg.sh check office1` is the one-command diagnosis (skill + routing row). + **>> TEST REBOOT FORCED by operator 2026-07-17** (prove it now, not wait for a natural one). The session + that issued `sudo reboot` died with the host. **POST-REBOOT SESSION -- verify BEFORE any manual `ip` + command:** (a) `systemctl status site-baseleg-office1.service` = active(exited), status=0; (b) its journal + (`sudo journalctl -u site-baseleg-office1.service -b`) logs **`added leg`** (leg was absent at boot -> the + unit CREATED it) -- `leg present` would mean something else persisted it, weaker proof; (c) `bash + scripts/site-baseleg.sh check office1` -> ok; (d) `ssh voffice1` returns live with NO manual step. All four + PASS -> RETIRE this carry-forward and record "boot-race PROVEN 2026-07-17". Any FAIL -> the ordering/poll + needs work (widen `BRIDGE_WAIT`, re-check which daemon creates the bridge) -- do NOT retire. 2. **DC leg rows.** `site-baseleg.sh` generalizes to the DCs but a DC row is DEFERRED-until-measured (hard rule 2; the harness fails on an un-commented DC supernet). At each DC standup, MEASURE the vcloud->DC reach and add a MEASURED leg row + install ONLY if that net is isolated (register item 20 / skill invariant).