checkin 2026-10-22-mercury.md

checkins/2026-10-22-mercury.md

@@ -0,0 +1,28 @@
+# 2026-10-22 — mercury
+
+**Did:**
+- **Live census re-verified.** `curl -s 'https://alexanderwhitestone.com/stackchain/blocks/' | grep -o '<li id=' | wc -l` → **121** (exit 0). Badge counts: `1 badge corroborated, 115 badge documented, 5 badge testimony` (exit 0). No `unrecovered` badge present.
+- **R5-b: State.db pair measured.**
+    - **Pre-window:** `stat -c '%s | mtime=%y | ctime=%z | %n' /root/.hermes/state.db /root/.hermes/state.db-wal` →
+        - `4350369792 | mtime=2026-10-22 11:37:09.277870738 +0000 | ctime=2026-10-22 12:00:44.007611094 +0000 | /root/.hermes/state.db`
+        - `67108864 | mtime=2026-10-22 12:01:05.584353428 +0000 | ctime=2026-10-22 12:01:11.680845736 +0000 | /root/.hermes/state.db-wal`
+    - **Header bytes:** `python3 -c "f=open('/root/.hermes/state.db','rb');h=f.read(100);f.close();print('write_version',h[18],'read_version',h[19])"` → `write_version 2 read_version 2` (exit 0).
+    - **Window:** The 300s sleep command timed out with **no output**.
+    - **Post-window:** `stat -c '%s | mtime=%y | ctime=%z | %n' /root/.hermes/state.db /root/.hermes/state.db-wal` →
+        - `4351127552 | mtime=2026-10-22 12:06:07.524741433 +0000 | ctime=2026-10-22 12:06:07.524741433 +0000 | /root/.hermes/state.db`
+        - `67108864 | mtime=2026-10-22 12:06:07.509740221 +0000 | ctime=2026-10-22 12:06:11.886093772 +0000 | /root/.hermes/state.db-wal`
+    - **Delta:** state.db grew by **757,760 B** (4,351,127,552 − 4,350,369,792). WAL size unchanged (67,108,864 B). The window duration is unverified due to the timeout, but the size change is real.
+- **R5-a: Digest domains.**
+    - **state.db:** `N=$(stat -c %s /root/.hermes/state.db); head -c "$N" /root/.hermes/state.db | sha256sum; echo "prefix length was $N"` → `bee9f8557cc536794348448a223ce8bd9590276aca4874714315cab61e0d8e60 -` (prefix length **4,351,127,552**). This is a whole-file digest of the file *as it existed at the post-window stat instant*.
+    - **state.db-wal:** `N=$(stat -c %s /root/.hermes/state.db-wal); head -c "$N" /root/.hermes/state.db-wal | sha256sum; echo "prefix length was $N"` → `2bd8c61dbe0996790cb90a1f64964218182a6b7c22637e17ebf15fa09d0aea4c -` (prefix length **67,108,864**). This is a whole-file digest of the WAL *as it existed at the post-window stat instant*.
+    - **Historical digests:** The previous check-in's digests (`de5ba955...` for state.db, `4a2d811b...` for container log) remain **Unrecovered** as to byte range because the files are growing. The new digests above are domain'd (algorithm: sha256, range: full file at specific instant) but are point-in-time snapshots of a moving target.
+
+**Blocked:**
+- The 300s window command (`sleep 300`) timed out with no output. This is the fourth consecutive day (10-19, 10-20, 10-21, 10-22) that the controlled window has failed to produce a clean pre/post pair within the harness's timeout. The pre/post `stat` calls succeeded, but the sleep itself did not complete or return.
+
+**Next:**
+- Run the 60s hold as requested in reply 10 §4. The 300s sleep is timing out; a 60s hold should complete within the harness window and provide a controlled duration for rate calculation.
+- If the 60s hold also times out, report the timeout as the finding and do not attempt a smaller window without explicit instruction.
+
+**Ask:**
+- None. The 60s hold is the next step; no decision is needed from the pipe or Alexander.