mirror of
https://github.com/bytedance/deer-flow.git
synced 2026-10-29 22:50:37 +00:00
13 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
Nan Gao
|
f39cdd6077
|
fix(mcp): keep rotating OAuth refresh tokens across restarts and Gateway instances (#7207)
* fix(mcp): keep rotating OAuth refresh tokens across restarts and Gateway instances
MCP servers on the OAuth refresh_token grant lost their rotated refresh
token whenever a new process or manager started (restart, tool-cache
reset for non-task servers, every personal-connection run, a second
Gateway instance) and re-seeded with the consumed configured token,
getting invalid_grant for good.
- New table mcp_oauth_tokens (migration 0047_mcp_oauth_tokens), one row
per refresh_token connection keyed (scope, owner_user_id, server_key):
('deployment', '', name) or ('personal', owner, personal runtime
name). Access and refresh tokens are DEER_FLOW_CREDENTIALS_KEY
ciphertext; a config fingerprint (token_url, client_id, grant, scope,
audience, seed) makes a changed seed win.
- OAuthTokenManager re-reads the row under its process lock, adopts a
peer's valid token, and otherwise claims the refresh with a lease plus
a version compare-and-set (database clock), fetches outside any
transaction, writes under the lease and releases it on failure.
- The Gateway installs the store (sync engine via to_thread) around the
LangGraph runtime for sqlite/postgres; personal managers pass their
owner. client_credentials stays in process; an unloadable key falls
back to process memory; database errors fail the refresh.
- credentials_key_consumers lists refresh_token MCP OAuth servers, so a
declared multi-instance deployment needs the shared key.
Docs: MCP_SERVER.md, CONFIGURATION.md, Helm README, README, mcp and
migrations AGENTS.md, CHANGELOG (en/zh).
* docs(changelog): link the MCP OAuth token store entry to #7207
* fix(mcp): prune removed connections' OAuth token rows and ask owners to re-authorize on invalid_grant
Review follow-up (willem-bd):
- Both MCP config writers drop mcp_oauth_tokens rows of connections a
save removes (deployment: _commit_mcp_config_write; personal:
_mutate, keeping each connection's enabled/disabled runtime-name
variants so a toggle keeps its chain). Best effort; no startup sweep,
since a missing config file would look like every server removed.
- invalid_grant on a refresh_token request raises
McpOAuthReauthorizationRequired: the tool error names the server and
asks for a new refresh token, plus the shared-key hint when this
instance could not decrypt the stored chain.
- Annotate the gateway helpers' config as AppConfig.
* fix(mcp): renew the OAuth refresh lease, keep unstored rotations, and key access tokens by request parameters
Review follow-up (willem-bd, second round):
- HTTPX timeouts bound each I/O wait, not a whole response: the refresh
holder now renews its lease every third of REFRESH_LEASE_SECONDS while
the token request runs and abandons it after REFRESH_MAX_SECONDS (120 s);
waiters outlast a live holder. Regression uses a real loopback endpoint
that streams its body for three lease lengths.
- A refresh whose tokens could not be stored keeps them pending
(unsaved_base_version): they are written back before an older stored
token is adopted or refreshed with, retried every few seconds while the
access token is still valid, and dropped once a peer stored a newer
chain. A failed save releases the lease.
- New token_request_fingerprint (extra_token_params and response fields):
a stored access token is reused only under the same request parameters;
otherwise the stored refresh-token chain is refreshed.
---------
Co-authored-by: Willem Jiang <willem.jiang@gmail.com>
|
||
|
Nan Gao
|
e8c7df9042
|
feat(channels): arbitrate IM transports with a per-provider leader lease (#6843)
* feat(channels): arbitrate IM transports with a per-provider leader lease Every Gateway instance connected every enabled IM platform with the same bot credentials, so channels could not run on more than one instance (Telegram 409 storms, Discord answering twice, scattered per-process state). channels.leader_lease was reserved by #6746; this implements it. - channel_leases table (migration 0041): one row per provider. Acquire is one dialect-native upsert decided against the database clock and bumps a fencing epoch; renew extends only a live lease at the holder's epoch; release sets the expiry to 0 so a peer takes over on its next tick. - app/channels/leadership.py: the holder's local monotonic deadline (send time of the last confirmed renewal + lease_seconds); holds() turns false one renewal interval early, so the transport stops at least grace_seconds before a peer may connect. DB calls are bounded. - ChannelService: start fence in _start_channel (no DB call) behind every start path; fixed-rate supervisor renews/acquires and runs transport starts/stops as per-provider tasks; takeover reloads the provider's saved config; failed starts back off; shutdown releases only leases whose transport stopped. get_status reports leader/leader_here and the holder's transport state on every instance. - Notification worker leaves rows of providers another instance holds. - Gate: the enabled=false refusal now recommends auto/true; an explicit true with database.backend memory is refused when a channel is enabled. - Docs, Helm, deployment guides, AGENTS.md and CHANGELOG updated. * docs(changelog): link the channel leader lease entry to #6843 * fix(channels): see peer-held leases for providers absent from local config Review on #6843: a provider connected from the web UI on a peer never enters this instance's channel config, so the notification worker here still claimed its rows (and parked them for 900s) and GET /api/channels/ reported no leader for it. Every supervision round, and startup, now also reads the whole lease table once (bounded), and notification exclusions and status use those observations: a provider whose live lease a peer holds is left to it even without local config. Running that read concurrently with renewals exposed that, through SQLAlchemy's aiosqlite adapter, an UPDATE ... RETURNING racing a read on another connection can hand back one of its own bind values instead of the column (stored rows stay correct; plain sqlite3 and raw aiosqlite do not show it). A lease must never take such a row for a confirmed renewal, so acquire/renew/release now decide by affected-row count and acquire reads its epoch back in the same transaction. --------- Co-authored-by: Willem Jiang <willem.jiang@gmail.com> |
||
|
Nan Gao
|
d18d2937e7
|
feat(channels): add channels.leader_lease and gate an explicitly disabled lease under multi-process signals (#6746)
* feat(channels): add channels.leader_lease and gate it under multi-process signals IM channels other than GitHub connect outbound to their platform. With more than one Gateway instance each one opens that connection with the same bot credentials: Telegram rejects the second poller with 409, Discord answers every message twice and per-process ChannelManager state diverges. The multi-process startup gate said nothing about channels; the only guard was a comment in the Helm values. Add the `channels.leader_lease` section (`enabled: auto|true|false`, `lease_seconds`, `grace_seconds`; startup-only, registered in the reload boundary) that the per-provider leader lease will read, validated by a Pydantic model from the non-schema `channels` mapping. The gate now refuses `enabled: false` under a multi-process signal while a non-GitHub provider is enabled under `channels.*` or connectable through `channel_connections`, naming the providers and the rollback; a malformed block is refused on every topology. `auto` follows the multi-process signal, so a single instance is unchanged. ChannelService strips the block before its per-provider loops and exposes it as `leader_lease`; `ensure_ready_channels` only visits registered providers. The two-Gateway harness overlay leaves the block alone. config_version 60 in the example, the Helm chart and its README; docs replace the comment-only "keep 1 replica" guidance with the gate. * docs(changelog): link #6746 * fix(channels): stop recommending the unimplemented leader lease as a remedy Review on #6746: the refusal told operators to switch channels.leader_lease from false to auto/true "so one instance owns each provider's connection", but the transport arbitration ships separately, so that change only bypassed the clause and every worker still started the same bot. The refusal now offers only the remedies that work today (a single instance, or disabling the channels) and states that auto/true are reserved for the lease that ships separately. A regression helper in the gate tests asserts the remedies start with the rollback and that auto/true are never offered. The config example, Helm values/README, CONFIGURATION.md, README, the deployment guide (en/zh), the model docstrings and both changelogs describe auto/true as reserved settings rather than an ownership guarantee. * fix(channels): warn on an enabled unknown channel key; reword the changelog entry Review on #6746: the registry guard in ensure_ready_channels skipped a misspelled provider key (channels.telegam) silently, where it previously reached _start_channel and logged 'Unknown channel type'. An enabled dict block that is neither a registered provider nor leader_lease now logs a warning that lists the registry. The changelog entry claimed the chart and docs drop the "keep 1 replica" guidance; they keep it (auto/true do not make several replicas safe yet) and only stop relying on the comment alone. * fix(dev): disable UI channel providers in multi-instance overlay --------- Co-authored-by: Willem Jiang <willem.jiang@gmail.com> |
||
|
Nan Gao
|
db7340d2fa
|
fix(security): add DEER_FLOW_CREDENTIALS_KEY and wire the channel credentials cipher (#6611)
* fix(security): add DEER_FLOW_CREDENTIALS_KEY for at-rest credential encryption
There was no at-rest encryption key. ChannelCredentialCipher had no
production caller: the channel service, the scheduled-task notification
outbox and the channel-connections router all built
ChannelConnectionRepository without a cipher, so store_credentials refused
(RuntimeError) and get_credentials always returned None.
- New harness module deerflow.config.credentials_key: env-only
DEER_FLOW_CREDENTIALS_KEY (comma-separated Fernet keys; first encrypts,
all decrypt via MultiFernet), CredentialsCipher with encrypt_text /
decrypt_text / rotate_text, a fernet:v2: prefix with legacy fernet:v1:
values still readable, and get_credentials_cipher() as the process-wide
entry point. Malformed keys are refused without echoing key material.
Without the variable a single instance generates
{base_dir}/.credentials_key. Never derived from AUTH_JWT_SECRET.
- New deerflow.config.secret_file.read_or_create_secret_file: O_EXCL
create (0600) plus read-back with a bounded settle window, shared by
.credentials_key, .jwt_secret (fixes two replicas cold-starting on a
shared volume keeping different JWT secrets) and the managed-model key.
- ChannelCredentialCipher now builds on CredentialsCipher (same
constructors); the Gateway loads the cipher off the event loop at
startup when channel_connections is enabled and passes it to every
ChannelConnectionRepository. Undecryptable values stay "missing + warn".
- Startup gate: a malformed key stops startup in any topology; a declared
multi-instance deployment with channel_connections enabled and no key
is refused. Plain multi-worker shares the generated file and is allowed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* feat(deploy): generate and inject DEER_FLOW_CREDENTIALS_KEY in Helm, compose and deploy.sh
- Helm: the app Secret generates a Fernet key (randBytes 32 translated to
the urlsafe alphabet) and preserves it across upgrades via lookup like
AUTH_JWT_SECRET; the gateway reads it through an optional secretKeyRef
so a user-managed existingAppSecret without it still boots (the Gateway
gate refuses only multi-instance deployments that store credentials).
README, NOTES and values document backup, rotation and existingAppSecret.
- scripts/deploy.sh resolves the key shell -> .env -> persisted file ->
generated, persisting it as $DEER_FLOW_HOME/.credentials_key (0600,
noclobber/O_EXCL with read-back): the same file the Gateway generates in
its runtime home, so either side's key is the one both keep.
- Both compose files pass DEER_FLOW_CREDENTIALS_KEY=${...:-} to the gateway.
- .env.example entry with a generation hint; README and CHANGELOG notes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(security): make abandoned secret-file replacement single-winner
A rename is last-writer-wins, so replicas that restart together and all
find the same abandoned empty secret file could each replace it and keep
a different value. The process that exclusively creates
<name>.replacing now owns the replacement, re-checks the file is still
empty, and publishes by renaming the claim over it; peers wait for that
value, bounded so a crashed replacer cannot block startup forever.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* docs(changelog): link DEER_FLOW_CREDENTIALS_KEY entries to #6611
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(security): name the setting that turns off a credentials-key consumer
The multi-instance refusal ended with "or disable
channel_connections.enabled=true"; it now says "or set
channel_connections.enabled=false".
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(security): never expose a secret file before its value is written
O_EXCL published the name before its content, so a creator that stalled
mid-write left an empty file a peer replaced after the settle window; the
resumed creator then returned a value nobody persisted, or deleted the
peer's valid file when its own write failed. Creators now write and sync
a private temporary file and hard-link it to the name, which fails if
the name exists; a failed creator only removes its own temporary file.
Filesystems without hard links publish through the same single-winner
<name>.replacing claim that replaces abandoned empty files, so creation
and replacement are serialized there.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(models): recover an empty managed-model key when no catalog exists
An interrupted first save could leave an empty key file; every later save
then passed empty bytes to Fernet and the endpoint returned 503 until the
file was removed by hand. An empty key with no catalog encrypted nothing,
so it now goes through the secret-file recovery path; next to an existing
catalog it is still refused.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
* fix(docker): pass the dotenv DEER_FLOW_CREDENTIALS_KEY through make docker-start
scripts/docker.sh runs Compose from docker/ without --env-file, so the dev
gateway's DEER_FLOW_CREDENTIALS_KEY=${DEER_FLOW_CREDENTIALS_KEY:-} entry
interpolated to empty and overrode the key the checkout .env supplied via
env_file. Export it from the dotenv like AUTH_TRUSTED_PROXIES.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
|
||
|
Nan Gao
|
b99ac916b6
|
fix(auth): count login lockouts in the database across Gateway replicas (#6501)
* fix(auth): count login lockouts in the database across Gateway replicas POST /api/v1/auth/login/local throttled failed logins per client IP in a module-level dict, so with N Gateway replicas behind one load balancer an attacker got N x auth.local.max_login_attempts guesses and a lockout on one replica was invisible to the others. Move the counter behind a LoginThrottleStore (deerflow.persistence. login_throttle): MemoryLoginThrottleStore keeps the historical per-process semantics and 10k tracked-IP bound; SqlLoginThrottleStore shares one login_throttle row per IP through the application database, incrementing and starting the lock in a single dialect-native upsert, applying check decisions with compare-and-set predicates in a separate write transaction, and sweeping served locks plus 24h-idle counters in bounded batches. Both stores pin the existing contract: the lock starts when the count reaches max_login_attempts, the committed duration is honored when the policy shrinks mid-lock, a served sentence is never resurrected, tightening keeps the count, success resets the IP, and the 429 status and message are unchanged. New startup-only selector auth.local.throttle_storage (auto | memory | db, default auto = the database whenever database.backend is sqlite or postgres). The Gateway installs the store once at startup after the persistence engine; deps._validate_login_throttle_storage warns when a multi-process deployment keeps memory. Migration 0034_login_throttle creates the table; config_version 55 -> 56 in config.example.yaml and the Helm chart. Docs: README, CONFIGURATION.md, AUTH_DESIGN.md, config and migrations AGENTS.md, CHANGELOG. * docs(changelog): link the login throttle entry to #6501 * fix(auth): re-check expiry on the row the login-throttle sweep deletes The bounded cleanup in SqlLoginThrottleStore.record_failure deleted `WHERE ip IN (SELECT ip ... WHERE <served or stale> LIMIT 200)`. On PostgreSQL READ COMMITTED the expiry predicate applied only inside the snapshot-bound subquery: when the DELETE had to wait on a row lock held by a concurrent check() that was extending the same lock under a raised lockout_seconds, PostgreSQL re-evaluated only the outer WHERE on the updated tuple, IP membership still held for the old candidate, and the now-active lock was deleted; the next attempt from that IP got a clean record and reached password verification (review of #6501, P2). Repeat the served/stale predicate on the DELETE target as well (SqlLoginThrottleStore.sweep_statement), keeping the bounded candidate subquery, so the re-evaluation sees the updated row's expiry. The other writes already carry their full compare-and-set predicate on the target row: _discard and the check() UPDATE both use _matches (ip, fail_count, locked_at, lock_duration_seconds), so a tuple changed while they waited no longer matches and the stale decision is skipped. _discard keeps the equality compare-and-set rather than adding an expiry condition: its early-release branch deletes a lock whose committed sentence has not elapsed under the lowered live policy, which an expiry predicate would wrongly block, and the equality predicate already pins every input of the decision. reset stays unconditional by design. Tests: a dialect-compile test asserts the DELETE's own WHERE carries the expiry predicate for postgresql and sqlite (red on the previous HEAD), and a PostgreSQL-only test (skipped without TEST_POSTGRES_URI) reproduces the update-versus-sweep interleaving with two engines, waiting on pg_stat_activity for the sweep to block before the extender commits. Docs: sql.py concurrency contract and the migrations guide's 0034 bullet. * fix(auth): keep an active login lock's clock and fail closed on explicit db Address the "changes requested" review on #6501. Lock clock (confirmed defect): record_failure re-stamped locked_at = now for every failure on an already-locked row, so a sustained attacker's sentence slid forward with each attempt. Both stores now apply one rule in one atomic decision: new_count = fail_count + 1; below max_attempts the IP counts with no lock (a failure after the operator raised the threshold clears an existing lock); at or over it, a failure during an *active* lock keeps locked_at and the committed lock_duration unchanged, while a never-locked or served row starts anew with now / lockout_seconds (so the failure that reaches a raised threshold starts a fresh lock rather than keeping a stale locked_at). The SQL upsert expresses it as nested case() on the row's own values; the memory store mirrors it. The sweeps in both stores now leave the row being recorded to that decision instead of deleting a served lock or stale counter first, which split the contract (SQL would have produced a fresh (1, NULL, NULL) row). base.py states the rule. Explicit db downgrade: resolve_login_throttle_store raises SystemExit — the failure type of the other startup validations in app.gateway.deps — when throttle_storage='db' resolves to the shared table but no engine is initialised, so langgraph_runtime refuses to start instead of silently running every replica on per-process counters; the lazy bare-app path applies the same rule and installs nothing. 'auto' keeps warning and falling back. 'db' on a memory database still degrades with a warning (there is no shared table to insist on). Non-blocking items: _key() logs a warning when a client key is truncated to 255 characters (never the key itself); the sql.py docstring covers wall-clock locked_at across replicas (run NTP; skew shifts evaluation by that much) and why STALE_COUNTER_SECONDS=24h / SWEEP_BATCH_SIZE=200 were chosen; a store-level blocking-I/O anchor covers the SQL reset path; the _validate_login_throttle_storage warning was verified to render 'throttle_storage=memory' for the StrEnum (the gate test now passes the enum member and asserts that text). reset stays unconditional: a peer failure that commits after a reset leaves fail_count=1, one genuine failure after the cleared history, which a new test documents. Docs: CHANGELOG entry, config.example.yaml, CONFIGURATION.md, AUTH_DESIGN.md and the throttle_storage field description. * fix(auth): re-evaluate a login-throttle check whose compare-and-set missed SqlLoginThrottleStore.check applied its decision (commit the live duration, or clear a served / released lock) with a compare-and-set predicate on the snapshot it had read, but ignored the affected-row count. When a peer changed the row between the read and the write — an admitted wrong password incrementing fail_count on another replica, or a successful login clearing the row — the UPDATE or DELETE matched zero rows, yet check reported the decision as applied: it could return an extended 60s sentence while the row kept its 1s duration, so the next check cleared the lock and let password verification through; a live duration decrease could be dropped the same way; and a row a peer had just cleared was still reported as locked (#6501 review, P2). Every compare-and-set write in check now checks result.rowcount. On a miss the row is re-read and re-evaluated from the top, up to CHECK_CAS_ATTEMPTS (3) times, applying a fresh compare-and-set decision on the new snapshot and never clobbering a reset or replacement: a row that is gone is reported as allowed, a fresh peer lock is reported as locked, and an incremented count gets the extension or decrease committed on it. Exhausting the bound fails closed: nothing is written and the answer is the last snapshot's lock under the live policy, logged at debug. The read stays in its own short transaction and each write in a separate one. The evaluation is now one shared pure function, evaluate_check in base.py (CheckDecision), used by both stores so the contract cannot diverge. The memory store keeps a single pass: its compare-and-set cannot miss because there is no await between the read and the write on the loop thread, which its docstring now states. Tests (red first on df26c2acf) use a session proxy that runs the peer mutation on a second engine right after the snapshot read ends its transaction: the reviewer's raised-duration interleaving (59.5s reported and committed, the increment survives, a later check still locked), the mirrored decrease, early release on the fresh row, a reset landing between read and write (0.0, no row recreated), and the bound-exhausted fail-closed case. * docs(changelog): resolve merge markers left in the Unreleased section * test(migrations): pin 0035_login_throttle as the head in the 0034 watermark test Upstream #6477 added test_migration_0034_run_event_seq_watermark.py asserting 0034 is the single head; this branch chains 0035_login_throttle after it, so the test now pins the current head and walks the chain like its siblings. * perf(auth): one session per login-throttle check, one transaction per failure Two non-blocking performance suggestions from the #6501 review. Single session per check. The router probed the row with store.get() and then called store.check(), which re-read it: a recorded IP cost two connection-pool checkouts and two SELECTs per request on the unauthenticated login path, letting a lockout flood compete with legitimate traffic for Postgres connections. LoginThrottleStore.check now takes the policy as an async callable (LoginThrottlePolicy) and resolves it lazily inside the store: the probe, the policy resolution and the compare-and-set loop share one session, a clean IP returns 0.0 without ever calling the policy (the config.yaml read is still skipped for the common case), and a recorded IP resolves it exactly once, also across compare-and-set retries. The decision is made on a snapshot read after the resolution: the policy read is a yield point, and a decision that needs no write has no compare-and-set to catch a row a peer changed meanwhile — so one session, but still a read after the yield. The memory store takes the same signature and re-reads its record after the await for the same reason. The router passes its asyncio.to_thread policy read as the callable and drops the separate probe; get stays on the protocol for tests and anchors. No short-TTL verdict cache: it would mask a successful login's reset for the TTL window, and the single session already halves the hot path. Sweep in the upsert's transaction, plus an index. record_failure ran the housekeeping DELETE in its own transaction and the upsert in a second one, two commits (two WAL flushes on PostgreSQL) per failed login; the sweep now rides in the upsert's transaction with one commit, the keep exclusion already protecting the row being recorded. The new ix_login_throttle_updated_at (ORM __table_args__ and the unreleased 0035 migration, created after the table and dropped before it, each guarded separately so a table left by an interrupted attempt still gets its index) serves the sweep's stale-counter predicate; the served-lock predicate scans a table the sweep itself keeps small. Tests (red first on a137fbc47): a clean IP never resolves the policy and a recorded IP resolves it exactly once (both stores), the SQL store opens exactly one session per check and commits once per record_failure (counting wrappers), the migration test asserts the index after upgrade, after completing a partial upgrade that lacked it, and gone after downgrade while compare_metadata stays empty; the router-level "mutated during policy resolution" tests now inject their mutation through the policy callable. * docs(changelog): note config_version 57 in the login throttle entry |
||
|
Nan Gao
|
5ae953fc2f
|
fix(memory): keep the retrieval index Pod-local and refresh scopes written by peers (#6494)
* fix(memory): keep the retrieval index Pod-local and refresh scopes written by peers
DeerMem's derived SQLite FTS5 index for every user was one WAL database pinned
to {storage_path}/.retrieval. With several Gateway Pods sharing one home volume
each Pod opened that file over a network filesystem (unsupported by SQLite),
every Pod start emptied and refilled it under its peers, one Pod's corruption
recovery deleted files the others held open, and a Pod never learned that a
peer had written a user's facts, so memory_search served stale results.
- New backend-private key memory.backend_config.retrieval_index_path: empty
keeps {storage_path}/.retrieval, a relative path resolves against
storage_path, an absolute path is used as-is. The full rebuild and the
corruption recovery act only on that configured index.
- rebuild_index records each agent scope's manifest signature
(mtime_ns, size, revision) before reading its facts; search_facts rebuilds a
scope whose live signature differs because another process wrote the user's
memory. An own commit advances the recorded signatures of that user's scopes
only when they were in sync at the previous revision, so own writes never
rebuild while an interleaved peer write still does.
- Gateway startup warns, in the style of _validate_agent_storage, when a
declared multi-instance deployment keeps the DeerMem index inside
storage_path.
- config.example.yaml documents the key; config_version 55 -> 56, mirrored in
the Helm chart and its README.
- Helm mounts an emptyDir at /var/lib/deerflow/memory-index, points the key at
it, drops the legacy memory.storage_path: memory.json line, and documents the
memory layout next to the other multi-replica prerequisites.
* docs(changelog): link the memory retrieval index entry to #6494
* fix(memory): never advance the index signature on a no-op commit and resync whole scopes
Two review findings on the cross-replica retrieval re-sync:
- save() and apply_changes() inferred the replaced manifest revision as
"returned revision - 1". The commit helper returns the existing manifest
unchanged for a no-op (unchanged-value or empty change set), so after a peer
bumped the manifest an own no-op write advanced this process's stale index
signature to the peer's revision without carrying the peer's delta: a fact the
peer deleted stayed searchable until the manifest changed again. Both paths
now read the real pre-commit revision under the user lock and pass
committed=(pre_revision, signature) only when the commit produced a new
revision; a no-op passes None so the next search still rebuilds the scope.
- The scoped rebuild (signature mismatch and the pre-existing dirty-scope path)
listed facts through list_facts() with its default limit of 100 and then
replaced every row of the scope, so a scope with more than 100 facts lost its
tail on the first re-sync. It now reads the scope's complete fact list.
* fix(memory): fence index signature promotion against a concurrent scope rebuild
Review finding on the cross-replica re-sync: a successful adapter notification
does not prove its delta is still in the index when the signature is promoted.
The SQLite lock covers single adapter calls only, and neither the scope lock
nor DeerMem's warm-up lock spans "notify, then promote", so rebuild_index can
replace a scope's rows in between. After a peer summary-only commit (r+1), A's
search starts a scoped refresh that captures the r+1 signature and reads f1/f2;
A's own deletion of f2 commits r+2 and removes the row; the refresh then
installs its stale f1/f2 and marks the scope synced at r+1; the deletion
resumes and promoted that signature to r+2, so later searches saw the live
signature as equal and kept returning the resurrected f2 until another manifest
change.
Mechanism: a generation fence rather than a lock. rebuild_index bumps a
per-scope generation (scoped rebuild) or a storage-wide generation (full
rebuild) under _cache_lock whenever it publishes or forgets a scope's rows, in
both mark_synced and forget_synced. _dispatch_retrieval_notifications snapshots
the generations under _cache_lock before its first adapter call and, at
promotion time, skips every scope whose generation moved (and everything when
the storage-wide generation moved), leaving the rebuild's own captured
signature in place; that signature predates the commit, so the next search's
comparison rebuilds the scope. A lock would have had to serialize notification
plus promotion with bulk install plus publication while search_facts ->
rebuild_index -> load() already takes the scope, file and cache locks; the
fence adds no lock-order obligations and only costs a redundant rebuild in the
rare race. Every interleaving is covered: a rebuild that completes before the
snapshot is reflected in the captured generation and promotion proceeds on an
index that carries both the rebuild's rows and this write's delta; one that
publishes between snapshot and promotion is skipped; one that publishes after
the promotion records its stale captured signature, which the next search
catches.
Tests reproduce both orders deterministically with threading.Event ordering
(adapter wrapper that holds the bulk install and the remove notification):
test_promotion_is_fenced_against_a_refresh_that_installs_in_between (red before
the fix) and
test_stale_refresh_installing_after_a_promotion_is_caught_by_the_next_search
(regression guard for the mirrored order).
* fix(memory): publish a rebuild's signature under the lock that replaced its rows
Review finding on the cross-replica re-sync: two peer-triggered refreshes of
one scope could replace rows in one order and publish their captured
signatures in the other. On a warmed Pod A, refresh R1 captures r+1 with f1/f2
and waits before installing; a peer deletes f2 at r+2; refresh R2 captures
r+2, installs f1 only and pauses after bulk_rebuild returns, before
mark_synced; R1 installs its stale f1/f2 and publishes r+1; R2 then publishes
r+2 over R1's rows. Later searches saw the live signature as equal and kept
returning the deleted fact. The generation fence only guards own-write
promotion, and the adapter's SQLite lock ends when the install returns.
Mechanism: a per-storage _retrieval_publish_lock held across the row
replacement and the signature publication, for the bulk path (bulk_rebuild +
mark_synced / forget_synced) and the clear+upsert fallback, scoped and full
alike. Overlapping refreshes of a scope therefore serialize their
install+publish pairs: whichever installs last also publishes last, and a stale
captured signature published last differs from the live manifest, so the next
search's compare rebuilds the scope. Fact reading (load() / the glob) stays
outside the lock so a slow read never blocks refreshes of other scopes. The
startup full rebuild holds the lock for its single bulk install; scoped
refreshes wait for it, which is acceptable because before warm-up completes a
search would be rebuilding its scope anyway (DeerMem.warm_retrieval takes its
own _retrieval_lock only after rebuild() returns, so the two never nest).
Lock order: the publish lock is acquired with no storage lock held --
search_facts releases _cache_lock before calling rebuild_index, reload() leaves
its scope/file-lock block before rebuilding, and DeerMem._ensure_retrieval_scopes
holds only its outer RLock, which nothing inside rebuild_index acquires. Inside
the lock only the adapter's own lock and _cache_lock are taken, and both are
leaf locks that never call back into storage, so no cycle is possible.
Test: test_rebuild_publication_stays_ordered_with_row_replacement drives the
reviewer's interleaving with threading.Event ordering (red before the fix). The
pause harness now holds a refresh after its fact read, outside the publish
lock, instead of inside the adapter's rebuild, so the earlier fence tests keep
their semantics without holding the new lock while paused.
* fix(memory): read each manifest once per full rebuild and serve the local index on a compare read error
Two suggestion-level review items on the cross-replica retrieval re-sync.
- The full rebuild cached manifest signatures keyed by (manifest path, agent
name), but _scope_signature reads only the manifest file -- every agent
bucket of a user shares one memory.json -- so the startup scan stat+read+
parsed the same file once per bucket that has facts. The cache is now keyed
by the manifest path alone while synced signatures are still recorded per
(user, agent) scope: users manifest reads instead of users x agents.
- The per-search freshness compare put a shared-volume manifest read on every
search_facts call, and _load_memory_file raises MemoryStorageCorruption on
OSError and malformed JSON, so a transient network-filesystem read error or a
corrupt memory.json failed a search the Pod-local index could still answer.
The compare now catches OSError, ValueError and MemoryStorageCorruption: it
logs a warning naming the scope (traceback at debug), serves the current
local index for that scope and leaves the recorded synced signature
untouched so the next search compares again. It deliberately does not mark
the scope dirty: a scoped rebuild would hit the same failing read, forget the
scope's signature and fail the search path too.
* docs(changelog): resolve merge markers left in the Unreleased section
* fix(memory): forget a scope's index signature when an own delta cannot be proved compatible
Review finding on the cross-replica re-sync: skipping promotion left a newer
rebuild's signature trusted even after this dispatcher had overwritten its rows
with an older notification. On a warmed Pod A, an own PATCH of f2 commits r+1
and pauses after releasing the user lock, before its adapter upsert; Pod B
deletes f2 at r+2; a concurrent search on A detects the peer change, rebuilds
the scope without f2 and publishes r+2; the pending r+1 upsert then reinserts
the deleted fact, and the revision check merely skipped because the synced
revision was r+2 rather than the PATCH's previous r. The signature stayed equal
to the live manifest, so searches returned the deleted fact indefinitely. The
generation fence could not help: it only suppresses promotion, the stale
adapter mutation still happens.
Rule: after an own commit's notifications are applied, the mutated scope's
recorded signature may stay only when the delta is provably compatible with
the published snapshot -- the scope is synced at exactly previous_revision and
no rebuild published rows for it (or storage-wide) since the fence snapshot --
in which case it is promoted as before. In every other case where this
dispatcher mutated the adapter for the scope (synced at another revision,
newer or older; generation moved; not synced at all; or notifications without
a committed revision, i.e. migration and clear paths) the scope's signature is
forgotten so the next search's compare rebuilds it. The user's other scopes
were not mutated by this delta: they are promoted when provable and otherwise
left untouched. The decision runs under _cache_lock together with the
generation reads, so a rebuild publishing concurrently is either observed or
observes the forget; the failure path (dirty + forget) is unchanged.
Orders covered: the reviewer's (publish r+2, then the stale upsert) forgets and
the next search rebuilds without f2; the earlier fence order (publish between
adapter op and promotion) now forgets too, one redundant rebuild; no concurrent
rebuild promotes exactly as before.
Test: test_stale_own_delta_replayed_over_a_newer_snapshot_forgets_the_scope
holds the PATCH's adapter upsert after the user lock is released (new
pause_next_upsert hook on the test adapter), Event-ordered, red before the fix.
---------
Co-authored-by: Willem Jiang <willem.jiang@gmail.com>
|
||
|
Nan Gao
|
5872d07e9a
|
fix(gateway): gate declared multi-instance deployments and require the redis stream bridge (#6328)
* fix(gateway): gate declared multi-instance deployments and require the redis stream bridge GATEWAY_WORKERS / WEB_CONCURRENCY only count the uvicorn workers of one process tree, so a Kubernetes Deployment with replicas > 1 (one worker per Pod) passed every multi-worker startup check while each Pod's startup orphan reconciliation wrote the other Pods' lease-less runs off as crashed on every rolling update. Add `deployment.multi_instance` as an explicit declaration that runs the same gate, plus a DEER_FLOW_MULTI_INSTANCE env override so deploy tooling can set the declaration from its replica count without editing config.yaml (wiring the Helm chart to export it is a separate change). Both the worker count and the declaration now also require a Redis stream bridge and refuse an explicit memory sandbox ownership store, since the memory bridge and ownership store are process-local. The agent-storage divergence warning, the inbound webhook dedupe warning and the WeChat QR-login guard honor the same declaration. * docs(changelog): link the multi-instance gate entry to #6328 * fix(ci): bump chart config_version to 53 and keep AGENTS.md under the size limits config.example.yaml moved to config_version 53 for the new deployment section, so the Helm chart's embedded config and its README example follow. The multi-instance notes in backend/AGENTS.md and backend/app/gateway/AGENTS.md shrink to one line each: main already sits within 55 bytes of the per-file hard limit for the gateway guide and within 163 bytes of the chain hard limit under backend/, and the full contract lives in backend/docs/CONFIGURATION.md. * fix(gateway): name every active knob in the multi-process rollback When GATEWAY_WORKERS > 1 and a multi-instance declaration are both active, the refusal offered only "Set GATEWAY_WORKERS=1", so an operator who followed it bounced through a second refusal on DEER_FLOW_MULTI_INSTANCE / deployment.multi_instance at the next start. The rollback now withdraws both knobs, and the browser-tools refusal takes the same path instead of the worker-count-only message. multi_instance_declaration() returns a frozen MultiInstanceDeclaration (source, value) with knob/rollback properties instead of a display string, so deps.py no longer branches on message prose. Tests pin the combined rollback text and that following it actually clears the gate. * docs(gateway): keep the gateway AGENTS.md under the size hard limit after merging main main grew backend/app/gateway/AGENTS.md by 30 bytes, which left it 25 bytes under the 49152-byte hard limit; the one-sentence multi-instance note from this branch pushed the merged file to 49164. Shorten that sentence by 18 bytes so scripts/check_agent_guidance.py passes on the merge result. * ci: re-run checks after test_stop_cancels_a_hung_driver_poll flaked on shard 4 tests/test_mcp_task_service.py::test_stop_cancels_a_hung_driver_poll timed out once on backend-unit-tests (4) for 2c16206ac. The MCP task service and its test are identical to main, main's run for cfe2023df passed, and 15 consecutive local runs pass; the test gives service.stop() a 1s budget while cancellation drain allows 5s, so a loaded runner can exceed it. Fork contributors cannot re-run a job, hence this empty commit. --------- Co-authored-by: Willem Jiang <willem.jiang@gmail.com> |
||
|
Grapette.L
|
fcd1d315de
|
fix(gateway): read uvicorn's WEB_CONCURRENCY in the multi-worker safety gates (#6252)
* fix(gateway): read uvicorn's WEB_CONCURRENCY in the multi-worker safety gates
The SQLite/scheduler gates in deps.py derived the worker count from
GATEWAY_WORKERS alone, but backend/Dockerfile and scripts/serve.sh launch
uvicorn without --workers, and uvicorn then takes the count from
WEB_CONCURRENCY. A Gateway scaled that way starts N processes with every
gate inert, which is exactly what the gates exist to prevent.
Read both variables, preferring GATEWAY_WORKERS, treat a blank value as
unset the way the compose command's ${GATEWAY_WORKERS:-1} does, and name
the variable that actually set the count in the refusal text. The channel
connection gate already reads both (channel_connections.py).
* fix(gateway): fall through to the fallback worker count
A non-numeric GATEWAY_WORKERS made the shared reader return one worker
immediately, so on the launches that pass no --workers (backend/Dockerfile,
scripts/serve.sh, backend/Makefile gateway) an operator's
GATEWAY_WORKERS=abc plus WEB_CONCURRENCY=4 started four processes while every
gate stayed inert. Keep looking at the remaining spellings and report a single
worker only after all of them have been tried.
The gate file also needs to be independent of the invoking shell: with
WEB_CONCURRENCY exported, the pre-existing
test_gate_treats_invalid_env_as_single_worker already failed on this branch,
because the reader now consults that variable too.
|
||
|
rain02333z-spec
|
828363705a
|
fix(scheduler): support safe multi-instance scheduler recovery (#4713)
* fix(scheduler): reject unsafe multi-worker startup * fix(scheduler): support safe multi-instance recovery * fix(scheduler): make multi-instance recovery lease-safe * fix(scheduler): address multi-instance review feedback * docs(scheduler): add multi-instance upgrade notes --------- Co-authored-by: rain02333z-spec <225106191+rain02333z-spec@users.noreply.github.com> |
||
|
Vanzeren
|
1c7531242c
|
feat(runtime): record terminal artifact delivery receipts (slice 1 of #4272) (#4365)
* feat(runtime): record terminal artifact delivery receipts (#4272) * fix(runtime): persist delivery receipts across recovery * test(runtime): cover delivery receipt invariants * fix(runtime): preserve terminal status on receipt outages |
||
|
Ryker_Feng
|
fa496c0c8d
|
feat(browser): add agentic browser control (#4187)
* feat(browser): add agentic browser control
* fix(frontend): format browser view changes
* fix(browser): keep browser optional and isolate sidecar layout
* fix(browser): address PR review security and IME findings
- Nginx: add a browser-stream WebSocket location before the generic
/api/threads regex so Live upgrades instead of downgrading to HTTP
(both nginx.conf and nginx.local.conf).
- Ownership: require an existing owned thread for the WS stream and REST
navigate, and tear down the browser session on thread deletion so a
later caller cannot reuse a retained page/cookies by guessing the id.
- SSRF: enforce the URL policy at the browser request boundary via a
context-level route guard covering redirects, popups, iframes, and
subresources (skipped for CDP-attached Chrome).
- IME: skip key forwarding while a composition is active so confirming a
CJK candidate with Enter no longer submits the remote page form.
Adds regression tests for the request guard, session teardown on delete,
and the composing-Enter key decision.
* fix(frontend): smooth streaming in long tool threads
* Revert "fix(frontend): smooth streaming in long tool threads"
This reverts commit f0462516eabe77f138d4027ea1c714fb226683cf.
* fix(browser): address review security and lifecycle findings
- Reject cross-origin WebSocket upgrades on the live browser stream
(Origin allow-list reuse of CORS/same-origin helpers) to close a
WS-CSRF hole, and fail closed when the ownership store is absent.
- Warn when a CDP-attached session runs with the SSRF request guard
off, and drop the unreachable CDP screencast teardown dead code.
- Read browser session launch config from a single canonical source
(browser_navigate) so it is deterministic regardless of call order.
- Bound per-thread Chromium accumulation with idle-timeout eviction
and an LRU max-sessions cap.
- Reset the Live reconnect counter on a successful open so the stream
can't permanently stall after the cumulative attempt cap.
* fix(frontend): reduce long tool thread render stalls
Reuse stable historical message groups during streaming, defer heavy Markdown and browser previews, and lazy-decode message images.
* fix(browser): keep live control responsive during continuous input
Why: Manual browser control felt laggy — a physical click ran the remote
Playwright click three times and each non-move input synchronously awaited a
JPEG screenshot, so events queued behind capture (queue wait up to ~237ms).
The first async attempt used a trailing-edge debounce, which froze the visible
page until a wheel/keyboard gesture stopped ("scroll finishes, then it jumps").
What:
- Frontend forwards one `click` per physical click instead of also emitting
`down`/`up`, so the remote page is not clicked twice per gesture.
- Backend detaches live-frame capture from input dispatch: non-move actions
start a rate-limited background refresh loop (leading frame + bounded cadence)
that keeps emitting frames while input continues and never blocks dispatch.
- Add regression tests: input dispatch no longer awaits the screenshot, rapid
inputs coalesce, and continuous input keeps refreshing before it stops.
Scenarios: Verified in the live Browser panel — a single click completes in
~57ms (was blocked behind a 171ms capture), and a 1.14s sustained wheel gesture
renders ~7 frames throughout the scroll instead of one frame after it ends.
* fix(browser): harden worker and session lifecycle
* fix(browser): address latest review feedback
* fix(frontend): preserve optimistic new-chat message
* test(e2e): preserve mocked message run ids
* fix(browser): address capability review feedback
---------
Co-authored-by: Willem Jiang <willem.jiang@gmail.com>
|
||
|
heart-scalpel
|
3bc3af2530
|
fix(runs): close multi-worker ownership gaps in run atomicity (#3948) (#4003)
* feat(runs): cross-process run ownership with lease + reconciliation (#3948) Implements work items 2 and 3 of the multi-worker P0 plan (docs/multi_worker.md). Work item 1 (Postgres startup gate, #3960) already landed; this PR makes run creation race-safe across worker processes and lets Postgres deployments recover orphaned inflight runs from crashed workers without mis-marking live runs as orphans. Work item 2 — cross-process atomic create_or_reject - Alembic revision 0004_run_ownership adds runs.owner_worker_id, runs.lease_expires_at, idx_runs_lease, and a partial unique index uq_runs_thread_active (one pending/running run per thread). The index is declared on RunRow.__table_args__ with sqlite_where + postgresql_where (mirroring uq_channel_connection_active_identity) so the empty-DB bootstrap path — which runs Base.metadata.create_all + alembic stamp head without executing any revision's upgrade() — also lands it on fresh deployments. Migration 0004 additionally creates it idempotently for legacy/versioned upgrades. - RunRepository.create_run_atomic is the new atomic primitive: - reject: INSERT directly; the partial unique index catches duplicate active runs; the manager surfaces the result as ConflictError. - interrupt/rollback: SELECT FOR UPDATE the conflicting rows, skip rows whose lease is still valid AND owned by another live worker (raise ConflictError — the INSERT would have failed on the index anyway, and a retry loop cannot make progress), cancel the rest in the same transaction, then INSERT the new row. Rows owned by this worker are interruptible regardless of lease state. - RunManager.create_or_reject dispatches to the store under the existing local lock; same-worker in-memory cancellation runs after the store commit succeeds. MemoryRunStore mirrors the same semantics for tests and database.backend=memory. Work item 3 — lease heartbeat + Postgres reconciliation - RunOwnershipConfig (lease_seconds=30, grace_seconds=10, heartbeat_enabled=false by default), registered as startup-only in reload_boundary.STARTUP_ONLY_FIELDS because the heartbeat background task is created once in langgraph_runtime() and is not rebuilt on config.yaml edits. - When heartbeat_enabled, each worker renewes leases on its own active runs with interval = lease_seconds / 3. The loop is bounded and stop-event-cancellable so shutdown is prompt. - reconcile_orphaned_inflight_runs now runs on every backend — the sqlite-only gate in app/gateway/deps.py is dropped in the same commit so there is no window where Postgres would mis-mark live Worker A runs as orphans. Reconciliation errors only runs whose lease is NULL (legacy pre-ownership rows) or older than grace_seconds. In single-worker mode (heartbeat off, NULL leases) all inflight rows reclaim immediately, preserving the pre-ownership recovery latency. - Heartbeat starts AFTER startup reconciliation and stops BEFORE the in-flight run drain on shutdown so the two cannot race. GATEWAY_WORKERS=1 with heartbeat_enabled=false keeps current behavior. Verified: 170 related tests + full backend suite (minus Docker-gated live tests) green; ruff check + ruff format clean. * fix(runs): tighten unique-violation handling and document clock-sync budget Three follow-up fixes to the cross-process run ownership work in #3948, surfacing during review. 1. _is_unique_violation: detect by driver-native signal, not message text The previous substring heuristic ("unique" + "violat", or "duplicate") missed SQLite's actual phrasing "UNIQUE constraint failed: <table>.<index>" — SQLite says "failed", not "violates", and never "duplicate". On SQLite the detector returned False, the reject path re-raised the raw IntegrityError, and clients saw HTTP 500 instead of ConflictError 409. The conversion is the load-bearing piece of the "store is source of truth" design but was untested — every atomic test used MemoryRunStore, which raises ConflictError directly and never reached this branch. Now prefers driver-native signals: psycopg pgcode/sqlcode "23505" and sqlite3 sqlite_errorcode SQLITE_CONSTRAINT_UNIQUE (reachable through SQLAlchemy IntegrityError.orig). Message matching stays as a fallback with SQLite's exact "unique constraint failed" phrase added. 2. interrupt/rollback: convert exhausted-retry IntegrityError to ConflictError The reject branch converts unique violations to ConflictError. The interrupt/rollback retry loop did not — on the 3rd attempt it re-raised the raw IntegrityError, leaking HTTP 500 for the same race condition that reject surfaces as 409. Symmetric conversion added after the loop; callers now see a consistent ConflictError regardless of strategy. 3. Document clock-sync requirement for multi-worker lease reconciliation reconcile_orphaned_inflight_runs compares another worker's UTC lease_expires_at against this worker's datetime.now(UTC). The only skew budget is grace_seconds (default 10s) — worst case, with the owning worker's heartbeat just about to fire, a peer whose clock is more than ~grace_seconds ahead can mis-reclaim a still-live run as an orphan. Documented in RunOwnershipConfig's docstring (with the math) and in config.example.yaml (with operational guidance), so operators in NTP-poor environments know to raise grace_seconds. Default unchanged: 10s is reasonable for NTP-synced K8s/cloud, and bumping it would slow recovery of genuinely dead workers (lease_seconds + grace_seconds from last heartbeat to reclaim). Tests: - test_create_run_atomic_reject_propagates_conflict_on_unique_violation: end-to-end against a real SQLite-backed RunRepository, pre-inserts an active run, asserts reject-strategy create surfaces as ConflictError rather than raw IntegrityError. - test_is_unique_violation_detects_real_sqlite_integrity_error: unit test for the detector against a real SQLite-raised IntegrityError; asserts driver-level sqlite_errorcode is SQLITE_CONSTRAINT_UNIQUE. - test_interrupt_exhausted_retries_surface_as_conflict_error: pins the symmetric 409 behavior after the retry loop exhausts. Verified: ruff check + ruff format clean; multi-worker + run_repository + owner_isolation + reload_boundary suites green. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(runs): close multi-worker ownership gaps in lease heartbeat and unique-violation detection Five code-review fixes from docs/multi_worker.md: 1. Drop unused ``claim_inflight_runs`` primitive — no caller anywhere. ``create_run_atomic`` does its own inline claim (SELECT FOR UPDATE + cancel) inside the INSERT transaction; a separate claim primitive would split that into two transactions and open a claim→INSERT race. Removes ~40 lines across base.py / memory.py / sql.py plus the unused ``now_iso`` parameter, freeing future RunStore implementations from providing it. 2. Broaden ``_renew_leases`` filter to renew pending/running runs owned by this worker even when ``record.task is None``. The previous ``task is not None`` requirement skipped the brief window between ``create_run_atomic`` inserting the row and the worker spawning the agent task; under event-loop load that window can approach ``lease_seconds``, after which peer reconciliation marks the run ``error`` (visible) or a peer's ``create_or_reject("interrupt")`` silently kills the queued run. Filter now: ``task is None or not task.done()``. 3. Document the unsynchronised ``record.lease_expires_at = new_expiry`` write. ``lease_expires_at`` is the only field on an existing record this path mutates; ``set_status`` / ``_persist_status`` touch other fields, so there is no concurrent writer to race against. Re-acquiring ``self._lock`` would serialise unrelated run mutations for no gain. 4. Gate ``_is_unique_violation`` message fallbacks on ``isinstance(current, (SAIntegrityError, sqlite3.IntegrityError))``. The driver-code path (pgcode/sqlite_errorcode) remains load-bearing; substring fallbacks are now belt-and-suspenders only for cases where the driver attribute isn't reachable through the cause chain. Without the gate, any application exception whose ``str()`` happens to contain "duplicate key" / "unique" + "violat" (CHECK constraint, validation error) would silently surface as HTTP 409 instead of 500. 5. Route ``update_lease`` through ``_call_store_with_retry`` for consistency with every other store call, and wrap ``await self._renew_leases()`` in ``_heartbeat_loop`` with ``except Exception: logger.warning(...)``. Previously a transient error from the snapshot path or an unexpected exception would kill the heartbeat task silently — after which no lease is ever renewed again and every active run eventually looks orphaned. ``except Exception`` lets ``CancelledError`` (BaseException since 3.8) propagate so shutdown cancellation still works. Regression tests: - ``test_heartbeat_renews_pending_run_before_task_is_spawned`` - ``test_is_unique_violation_does_not_misclassify_application_exception`` Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(runs): harden multi-worker migration, memory atomicity, and tz-naive lease comparison Three follow-up fixes to the multi-worker run ownership work: - migration 0004 dedupe pass: cancel superseded duplicate active rows per thread before creating the partial UNIQUE index ``uq_runs_thread_active`` so dirty DBs (Postgres deployments that had reconciliation skipped by the old sqlite-only gate, or any env that ran GATEWAY_WORKERS>1 before this PR) do not abort the alembic upgrade and block gateway startup. Keeps the newest active row per thread, marks the rest as error with an explanatory message. - MemoryRunStore.create_run_atomic interrupt/rollback path: split the single- pass loop into two passes (collect candidates, validate, then mutate) so a ConflictError raised on a later candidate does not leave earlier candidates half-interrupted. Mirrors the SQL store's transactional rollback semantics; the entire test_multi_worker_run_ownership.py suite runs against memory so this divergence was giving false confidence. - RunRepository.create_run_atomic interrupt path: coerce tz-naive ``row.lease_expires_at`` to UTC before comparing against the aware ``cutoff``. SQLite drops tzinfo on read despite ``DateTime(timezone=True)`` (this file's own comment acknowledges it), so the Python-side comparison raised ``TypeError: can't compare offset-naive and offset-aware datetimes`` whenever heartbeat was enabled on SQLite and a lease was non-NULL. Defaults (heartbeat off -> leases always NULL) masked it, but there was no guard against the combination. Follows the existing "naive is UTC" convention from ``coerce_iso``. Each fix ships with a regression test pinning the behavior. Co-Authored-By: heart-scalpel <heart-scalpel@users.noreply.github.com> Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(runs): enforce heartbeat for multi-worker, fix memory-store datetime comparison, lazy-import ConflictError in store layer Three fixes from code review: 1. Extend the startup gate (GATEWAY_WORKERS>1) to also require run_ownership.heartbeat_enabled=true. Without heartbeat every run has a NULL lease, so reconciliation treats all inflight rows as orphans and Worker B would kill Worker A's live runs on every rolling update or scale-up. 2. Fix MemoryRunStore.list_inflight_with_expired_lease to parse created_at as datetime instead of ISO string lexical comparison, and handle tz-naive lease values uniformly with the SQL store. 3. Store layer (sql.py, memory.py) now lazy-imports ConflictError inside create_run_atomic instead of importing from the higher RunManager layer at module level. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(runs): add owner check to update_lease, document create() assumption, restore deleted comment - update_lease (SQL + memory) now requires owner_worker_id match in WHERE clause so the primitive is safe by construction against misuse - create() docstring notes it bypasses atomic create_run_atomic and assumes no active run exists for the thread - restore explanatory comment in MemoryRunStore.aggregate_tokens_by_thread that was dropped in an earlier commit Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(runs): add psycopg3 sqlstate detection and periodic orphan reconciliation - _is_unique_violation now checks sqlstate attribute (psycopg3 uses this instead of pgcode). On Postgres, the only supported multi-worker backend, detection was falling through to the message-substring fallback. - _heartbeat_loop now runs reconcile_orphaned_inflight_runs every 3rd cycle (every lease_seconds) to catch orphans whose lease expires between pod restarts. Single-worker deployments are unaffected (heartbeat off). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com> Co-authored-by: heart-scalpel <heart-scalpel@users.noreply.github.com> Co-authored-by: Willem Jiang <willem.jiang@gmail.com> |
||
|
heart-scalpel
|
8cde7f258e
|
feat(gateway): refuse multi-worker startup on non-Postgres backends (#3960)
Why: issue #3948 identifies four correctness breakers when GATEWAY_WORKERS > 1. Work item 1 adds a startup gate that refuses to boot when database.backend is not postgres, giving operators a clear error instead of silent SQLite write-lock corruption. The gate runs inside langgraph_runtime() before init_engine_from_config, so a misconfigured deploy never opens a listener or writes to disk. Non-integer env values fall back to 1 so uvicorn's own validation is unaffected. |