findings: record #49829 (Defect 3 filed upstream)

Fourth upstream issue filed: HIT_PENDING no-deadline hang in the shared
tiering manager, Codex-reviewed. All four known crash/stall defects are
now filed; #49820 and #49829 remain unfixed upstream.

Signed-off-by: nilig <nili.ifergan@gmail.com>

Author: nilig

test/p2p-findings/upstream-issues-map.md

@@ -1,25 +1,27 @@
 # Upstream vLLM P2P issues: map + coverage
 
-Maps the three filed upstream issues to the two internal defect catalogs
+Maps the four filed upstream issues to the two internal defect catalogs
 (`p2p-pd-defects.md`, `p2p-lookup-hangs.md`) and states what remains after
-each upstream fix lands. All three issues are OPEN as of 2026-07-25.
+each upstream fix lands. All four issues are OPEN as of 2026-07-25.
 
-## The three issues (distinct bugs, distinct layers)
+## The four issues (distinct bugs, distinct layers)
 
 | issue | what it is | layer / file | crash or stall | catalog | fix PR |
 |---|---|---|---|---|---|
 | #49635 | late KV-offload store after request finalization -> `KeyError` (`_req_state` deleted at `tiering/manager.py:721`, indexed unguarded at `:542` via `scheduler.py:1045 _build_store_jobs`) | OffloadingConnector store path -- NOT P2P-specific | crash | neither catalog | #49671 (open) |
 | #49809 | reconnect to a reaped peer trips `AssertionError: ZmqConnection already exists`; dead conn never released | P2P control transport (`tiering/p2p/control/zmq.py`) | crash | pd-defects Defect 2 | #49823 (open) |
 | #49820 | symmetric producer accepts a fetch it cannot serve, never sends `TransferDone(success=False)` -> consumer deferred full `_LOAD_TIMEOUT_S=30s` | P2P session (`tiering/p2p/session/*`) | stall | new (post-Liran residual) | none |
+| #49829 | `TieringOffloadingManager.lookup()` returns `HIT_PENDING` unconditionally, no deadline, no downgrade-to-MISS | shared tiering manager (`tiering/manager.py`) -- NOT P2P-specific | stall | lookup-hangs Defect 3 | none |
 
-Distinguishing axis: #49635 is OffloadingConnector-general (fires even
-aggregated / no P2P); #49809 and #49820 are P2P-tier-specific but different
-sub-layers (control transport vs session).
+Distinguishing axis: #49635 and #49829 are OffloadingConnector-general (fire
+even aggregated / no P2P, different files from each other); #49809 and #49820
+are P2P-tier-specific but different sub-layers (control transport vs session).
 
-## Coverage: are those three enough?
+## Coverage: are those four enough?
 
-No. After #49635 (->#49671), #49809 (->#49823), #49820 (->future fix), and
-Liran's #48021 all land:
+Yes, this is now the complete known set. After #49635 (->#49671), #49809
+(->#49823), #49820 (->future fix), #49829 (->future fix), and Liran's #48021
+all land:
 
 | defect | covered by | patch still needed? |
 |---|---|---|
@@ -28,7 +30,7 @@ Liran's #48021 all land:
 | #49820 symmetric-fetch 30s stall | #49820 fix (TBD) | no, once written |
 | lookup-hangs Defect 1 (duplicate-fetch teardown) | Liran's one-fetch contract in #48021 (`dupfetch=0` on a version-matched engine) | no |
 | lookup-hangs Defect 2 (pop-on-read MISS livelock) | Liran's `register_lookup` retains resolved HIT+MISS until fetch/finish (no pop-on-read) = the sticky-MISS. VERIFIED in `client.py` @145a460c | no (`defect12` Part 2 redundant) |
-| lookup-hangs Defect 3 (`HIT_PENDING` write-in-flight, no deadline in UPSTREAM `tiering/manager.py`) | nothing | YES -- `defect3_fix_pending-wait-deadline.diff`, or file upstream |
+| lookup-hangs Defect 3 = #49829 (`HIT_PENDING` write-in-flight, no deadline in `tiering/manager.py`) | #49829 fix (TBD) | no, once written -- `defect3_fix_pending-wait-deadline.diff` validates the direction (deterministic repro: `repro_defect3.py`) |
 
 Softer gap: `defect12` Part 1 (lookup-phase deadline) is NOT in Liran's branch
 -- only `_LOAD_TIMEOUT_S=30` on the fetch phase, no deadline on a stranded
@@ -46,14 +48,16 @@ Two catalogs both start at "Defect 1", causing collisions:
 
 ## Net
 
-Two crashes fixed (#49635, #49809), the 30s stall fixed (#49820), and
-lookup-hangs Defect 1+2 covered by #48021. The remaining real bug is
-lookup-hangs Defect 3 -- an upstream `tiering/manager.py` hang untouched by any
-P2P issue.
+Two crashes fixed (#49635 -> #49671, #49809 -> #49823), and lookup-hangs
+Defect 1+2 covered by #48021. Two stalls remain unfixed: #49820 (P2P session,
+symmetric fetch) and #49829 (shared tiering manager, `HIT_PENDING`
+write-in-flight -- lookup-hangs Defect 3). Both have a documented mechanism, a
+deterministic repro, and a validated local patch; neither has an upstream fix
+yet.
 
 VERIFIED 2026-07-25 against `main`@70009fb9: `TieringManager.lookup()` still
 returns `LookupResult.HIT_PENDING` bare, with no deadline. Upstream added only
 observability around it (`secondary_lookup_start_time`, `sync_lookup_delay`
-accumulation) -- no bound, no downgrade-to-MISS. So Defect 3 is LIVE in main;
-`defect3_fix_pending-wait-deadline.diff` is still required. Not filed upstream
--> candidate for a companion issue to #49809/#49820.
+accumulation) -- no bound, no downgrade-to-MISS. So Defect 3 was LIVE in main;
+filed as **#49829** (Codex-reviewed, ASCII-only, cc @orozery @ronensc).
+`defect3_fix_pending-wait-deadline.diff` is still required until it lands.