logos-delivery/tests/persistency/test_facade.nim
NagyZoltanPeter e8566db8ee
fix(persistency): own Persistency per node instead of a process-global singleton (#4109)
* fix(persistency): own Persistency per node instead of a process-global singleton

The Persistency singleton (gPersistency) was allocated on whichever FFI
thread first ran waku.start; under --mm:refc its memory belonged to that
thread's heap, so a second library context adopting it read a foreign
heap (SIGSEGV in sdsPersistence -> openJob -> tables.rawGet), one
context's stop stole the other's SDS persistence, and a destroyed
context poisoned re-init with a different local-storage-path.

- remove the singleton (instance/reset/gPersistency); Persistency.new is
  the public constructor, no lock needed (instances are thread-confined)
- own the instance as Waku.persistency: created and provided in
  waku.start, cleared and closed in waku.stop on the owning thread
- expose it via a sync GetPersistency RequestBroker scoped to the node's
  BrokerContext; sdsPersistence resolves through it (same-thread ref
  return, no marshalling)
- add InMemoryStoragePath (":memory:") support: private in-memory SQLite
  per job worker, for tests
- rewrite test_singleton as per-instance + broker coverage; rewrite
  test_thread_affinity from a known-failing UB repro into a regression
  guard (two in-memory jobs, worker spinup/teardown, cross-thread broker
  denial) and register it in test_all; migrate remaining tests to
  new/close; the FFI lifecycle test's stop-steals and different-paths
  cases now pass against the real dylib

The FFI destroy-without-stop teardown gap remains tracked in #4108.

Fixes #4103

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(persistency): create the instance only after startup fully succeeds

waku.start has many error return paths; creating Persistency early meant
every one of them left the field set and the GetPersistency provider
installed with no teardown. Persistency.new is inert (no threads or
files until the first openJob) and every consumer runs post-start, so
creating and providing it as the last startup step removes the need for
any error-path cleanup entirely.

Addresses PR #4109 review feedback.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* fix(persistency): create early for startup-stage restores, tear down on failed start

Creating the instance as the last startup step made it impossible for
any stage of start to restore persisted data (e.g. a future store-state
restore). Restore the original ordering -- create and provide the
instance first -- and cover every error return path of waku.start with a
success-flag defer that clears the provider and closes the instance.
Teardown is factored into closePersistency, shared by stop and the
failed-start path.

Addresses PR #4109 review discussion.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* chore(persistency): address remaining Copilot review findings

- document that Persistency instances are thread-confined (not
  thread-safe) on the type itself, pointing at the GetPersistency broker
  as the sanctioned access path
- use tryRemoveFile for the FFI test's log cleanup so an unremovable
  file cannot fail the test for unrelated reasons

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

* Fix comment

---------

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-08-06 23:35:10 +02:00

197 lines
5.3 KiB
Nim

{.used.}
import std/[os, strutils, times]
import chronos, results
import testutils/unittests
import logos_delivery/waku/persistency/persistency
proc payload(s: string): seq[byte] =
result = newSeq[byte](s.len)
for i, c in s:
result[i] = byte(c)
template str(b: seq[byte]): string =
var s = newString(b.len)
for i, x in b:
s[i] = char(x)
s
proc tmpRoot(label: string): string =
let p = getTempDir() / ("persistency_facade_" & label & "_" & $epochTime().int)
removeDir(p)
p
# Bounded poll on exists() to bridge the documented persist->read race.
proc waitUntilExists(
t: Job, category: string, k: Key, timeoutMs = 1000
): Future[bool] {.async.} =
let deadline = epochTime() + (timeoutMs.float / 1000.0)
while epochTime() < deadline:
let r = await t.exists(category, k)
if r.isOk and r.get():
return true
await sleepAsync(chronos.milliseconds(2))
return false
suite "Persistency facade":
asyncTest "persistPut then get round-trips":
let root = tmpRoot("put_get")
defer:
removeDir(root)
let p = Persistency.new(root).get()
defer:
p.close()
let t = p.openJob("t").get()
let k = key("c", 1'i64)
await t.persistPut("msg", k, payload("hi"))
let ckOk1 = await t.waitUntilExists("msg", k)
check ckOk1
let aw1 = await t.get("msg", k)
let got = aw1.get()
check got.isSome
check str(got.get) == "hi"
asyncTest "persist (batch) is atomic and visible together":
let root = tmpRoot("batch")
defer:
removeDir(root)
let p = Persistency.new(root).get()
defer:
p.close()
let t = p.openJob("t").get()
var ops: seq[TxOp]
for i in 1'i64 .. 4:
ops.add(
TxOp(category: "msg", key: key("c", i), kind: txPut, payload: payload($i))
)
await t.persist(ops)
let ckOk2 = await t.waitUntilExists("msg", key("c", 4'i64))
check ckOk2
let aw2 = await t.count("msg", prefixRange(key("c")))
let cnt = aw2.get()
check cnt == 4
asyncTest "scanPrefix returns rows in key order":
let root = tmpRoot("scan")
defer:
removeDir(root)
let p = Persistency.new(root).get()
defer:
p.close()
let t = p.openJob("t").get()
for i in [3'i64, 1, 4, 1, 5, 9, 2]:
await t.persistPut("msg", key("c", i), payload($i))
let ckOk3 = await t.waitUntilExists("msg", key("c", 9'i64))
check ckOk3
let aw3 = await t.scanPrefix("msg", key("c"))
let rows = aw3.get()
# 7 ops with duplicate key i=1 -> 6 distinct rows
check rows.len == 6
var seenOrder: seq[int]
for r in rows:
seenOrder.add(parseInt(str(r.payload)))
check seenOrder == @[1, 2, 3, 4, 5, 9]
asyncTest "scanPrefix reverse=true returns rows in reverse order":
let root = tmpRoot("scan_rev")
defer:
removeDir(root)
let p = Persistency.new(root).get()
defer:
p.close()
let t = p.openJob("t").get()
for i in 1'i64 .. 3:
await t.persistPut("msg", key("c", i), payload($i))
let ckOk4 = await t.waitUntilExists("msg", key("c", 3'i64))
check ckOk4
let aw4 = await t.scanPrefix("msg", key("c"), reverse = true)
let rows = aw4.get()
check rows.len == 3
check str(rows[0].payload) == "3"
check str(rows[2].payload) == "1"
asyncTest "deleteAcked round-trips and reports row presence":
let root = tmpRoot("delete")
defer:
removeDir(root)
let p = Persistency.new(root).get()
defer:
p.close()
let t = p.openJob("t").get()
let k = key("c", 1'i64)
let aw5 = await t.deleteAcked("msg", k)
let miss = aw5.get()
check miss == false
await t.persistPut("msg", k, payload("v"))
let ckOk5 = await t.waitUntilExists("msg", k)
check ckOk5
let aw6 = await t.deleteAcked("msg", k)
let hit = aw6.get()
check hit == true
let aw7 = await t.exists("msg", k)
check aw7.get() == false
asyncTest "persistDelete fire-and-forget removes the row":
let root = tmpRoot("fadel")
defer:
removeDir(root)
let p = Persistency.new(root).get()
defer:
p.close()
let t = p.openJob("t").get()
let k = key("c", 1'i64)
await t.persistPut("msg", k, payload("v"))
let ckOk6 = await t.waitUntilExists("msg", k)
check ckOk6
await t.persistDelete("msg", k)
# Poll for absence.
let deadline = epochTime() + 1.0
var gone = false
while epochTime() < deadline:
let aw8 = await t.exists("msg", k)
if not aw8.get():
gone = true
break
await sleepAsync(chronos.milliseconds(2))
check gone
asyncTest "two jobs do not see each other's data via the facade":
let root = tmpRoot("iso")
defer:
removeDir(root)
let p = Persistency.new(root).get()
defer:
p.close()
let a = p.openJob("a").get()
let b = p.openJob("b").get()
let k = key("c", 1'i64)
await a.persistPut("msg", k, payload("A"))
await b.persistPut("msg", k, payload("B"))
let ckOk7 = await a.waitUntilExists("msg", k)
check ckOk7
let ckOk8 = await b.waitUntilExists("msg", k)
check ckOk8
let aw9 = await a.get("msg", k)
check str(aw9.get().get) == "A"
let aw10 = await b.get("msg", k)
check str(aw10.get().get) == "B"
let aw11 = await a.count("msg", prefixRange(key("c")))
check aw11.get() == 1
let aw12 = await b.count("msg", prefixRange(key("c")))
check aw12.get() == 1