feat[marketplace]: add slot queue pausing (#752)

* add seen flag

* Add MockSlotQueueItem and better prioritisation tests

* Update seen priority, and include in SlotQueueItem.init

* Re-add processed slots to queue

Re-add processed slots to queue if the sale was ignored or errored

* add pausing of queue

- when processing slots in queue, pause queue if item was marked seen
- if availability size is increased, trigger onAvailabilityAdded callback
- in sales, on availability added, clear 'seen' flags, then unpause the queue
- when items pushed to the queue, unpause the queue

* remove unused NoMatchingAvailabilityError from slotqueue

The slot queue should also have nothing to do with availabilities

* when all availabilities are empty, pause the queue

An empty availability is defined as size < DefaultBlockSize as this means even the smallest possible request could not be served. However, this is up for discussion.

* remove availability from onAvailabilitiesEmptied callback

* refactor onAvailabilityAdded and onAvailabilitiesEmptied

onAvailabilityAdded and onAvailabilitiesEmptied are now only called from reservations.update (and eventually reservations.delete once implemented).

- Add empty routine for Availability and Reservation
- Add allEmpty routine for Availability and Reservation, which returns true when all all Availability or Reservation objects in the datastore are empty.

* SlotQueue test support updates

* Sales module test support updates

* Reservations module tests for queue pausing

* Sales module tests for queue pausing

Includes tests for sales states cancelled, errored, ignored to ensure onCleanUp is called with correct parameters

* SlotQueue module tests for queue pausing

* fix existing sales test

* PR feedback

- indent `self.unpause`
- update comment for `clearSeenFlags`

* reprocessSlot in SaleErrored only when coming from downloading

* remove pausing of queue when availabilities are "emptied"

Queue pausing when all availiabilies are "emptied" is not necessary, given that the node would not be able to service slots once all its availabilities' freeSize are too small for the slots in the queue, and would then be paused anyway.

Add test that asserts the queue is paused once the freeSpace of availabilities drops too low to fill slots in the queue.

* Update clearing of seen flags

The asyncheapqueue update overload would need to check index bounds and ultimately a different solution was found using the mitems iterator.

* fix test

request.id was different before updating request.ask.slots, and that id was used to set the state in mockmarket.

* Change filled/cleanup future to nil, so no await is needed

* add wait to allow items to be added to queue

* do not unpause queue when seen items are pushed

* re-add seen item back to queue once paused

Previously, when a seen item was processed, it was first popped off the queue, then the queue was paused waiting to process that item once the queue was unpaused. Now, when a seen item is processed, it is popped off the queue, the queue is paused, then the item is re-added to the queue and the queue will wait until unpaused before it will continue popping items off the queue. If the item was not re-added to the queue, it would have been processed immediately once unpaused, however there may have been other items with higher priority pushed to the queue in the meantime. The queue would not be unpaused if those added items were already seen. In particular, this may happen when ignored items due to lack of availability are re-added to a paused queue. Those ignored items will likely have a higher priority than the item that was just seen (due to it having been processed first), causing the queue to the be paused.

* address PR comments

codex/sales.nim

@@ -78,13 +78,13 @@ proc onProve*(sales: Sales): ?OnProve = sales.context.onProve
 
 proc onExpiryUpdate*(sales: Sales): ?OnExpiryUpdate = sales.context.onExpiryUpdate
 
-func new*(_: type Sales,
+proc new*(_: type Sales,
           market: Market,
           clock: Clock,
           repo: RepoStore): Sales =
   Sales.new(market, clock, repo, 0)
 
-func new*(_: type Sales,
+proc new*(_: type Sales,
           market: Market,
           clock: Clock,
           repo: RepoStore,
@@ -111,16 +111,20 @@ proc remove(sales: Sales, agent: SalesAgent) {.async.} =
 proc cleanUp(sales: Sales,
              agent: SalesAgent,
              returnBytes: bool,
+             reprocessSlot: bool,
              processing: Future[void]) {.async.} =
 
   let data = agent.data
 
-  trace "cleaning up sales agent",
+  logScope:
-    requestId = data.requestId,
+    topics = "sales cleanUp"
-    slotIndex = data.slotIndex,
+    requestId = data.requestId
-    reservationId = data.reservation.?id |? ReservationId.default,
+    slotIndex = data.slotIndex
+    reservationId = data.reservation.?id |? ReservationId.default
     availabilityId = data.reservation.?availabilityId |? AvailabilityId.default
 
+  trace "cleaning up sales agent"
+
   # if reservation for the SalesAgent was not created, then it means
   # that the cleanUp was called before the sales process really started, so
   # there are not really any bytes to be returned
@@ -132,7 +136,6 @@ proc cleanUp(sales: Sales,
                       )).errorOption:
           error "failure returning bytes",
             error = returnErr.msg,
-            availabilityId = reservation.availabilityId,
             bytes = request.ask.slotSize
 
   # delete reservation and return reservation bytes back to the availability
@@ -141,10 +144,21 @@ proc cleanUp(sales: Sales,
                     reservation.id,
                     reservation.availabilityId
                   )).errorOption:
-      error "failure deleting reservation",
+      error "failure deleting reservation", error = deleteErr.msg
-        error = deleteErr.msg,
+
-        reservationId = reservation.id,
+  # Re-add items back into the queue to prevent small availabilities from
-        availabilityId = reservation.availabilityId
+  # draining the queue. Seen items will be ordered last.
+  if reprocessSlot and request =? data.request:
+    let queue = sales.context.slotQueue
+    var seenItem = SlotQueueItem.init(data.requestId,
+                                      data.slotIndex.truncate(uint16),
+                                      data.ask,
+                                      request.expiry,
+                                      seen = true)
+    trace "pushing ignored item to queue, marked as seen"
+    if err =? queue.push(seenItem).errorOption:
+      error "failed to readd slot to queue",
+        errorType = $(type err), error = err.msg
 
   await sales.remove(agent)
 
@@ -176,8 +190,8 @@ proc processSlot(sales: Sales, item: SlotQueueItem, done: Future[void]) =
     none StorageRequest
   )
 
-  agent.onCleanUp = proc (returnBytes = false) {.async.} =
+  agent.onCleanUp = proc (returnBytes = false, reprocessSlot = false) {.async.} =
-    await sales.cleanUp(agent, returnBytes, done)
+    await sales.cleanUp(agent, returnBytes, reprocessSlot, done)
 
   agent.onFilled = some proc(request: StorageRequest, slotIndex: UInt256) =
     sales.filled(request, slotIndex, done)
@@ -222,7 +236,6 @@ proc mySlots*(sales: Sales): Future[seq[Slot]] {.async.} =
   return slots
 
 proc activeSale*(sales: Sales, slotId: SlotId): Future[?SalesAgent] {.async.} =
-  let market = sales.context.market
   for agent in sales.agents:
     if slotId(agent.data.requestId, agent.data.slotIndex) == slotId:
       return some agent
@@ -241,60 +254,29 @@ proc load*(sales: Sales) {.async.} =
       slot.slotIndex,
       some slot.request)
 
-    agent.onCleanUp = proc(returnBytes = false) {.async.} =
+    agent.onCleanUp = proc(returnBytes = false, reprocessSlot = false) {.async.} =
-      let done = newFuture[void]("onCleanUp_Dummy")
+      # since workers are not being dispatched, this future has not been created
-      await sales.cleanUp(agent, returnBytes, done)
+      # by a worker. Create a dummy one here so we can call sales.cleanUp
-      await done # completed in sales.cleanUp
+      let done: Future[void] = nil
+      await sales.cleanUp(agent, returnBytes, reprocessSlot, done)
+
+    # There is no need to assign agent.onFilled as slots loaded from `mySlots`
+    # are inherently already filled and so assigning agent.onFilled would be
+    # superfluous.
 
     agent.start(SaleUnknown())
     sales.agents.add agent
 
 proc onAvailabilityAdded(sales: Sales, availability: Availability) {.async.} =
-  ## Query last 256 blocks for new requests, adding them to the queue. `push`
+  ## When availabilities are modified or added, the queue should be unpaused if
-  ## checks for availability before adding to the queue. If processed, the
+  ## it was paused and any slots in the queue should have their `seen` flag
-  ## sales agent will check if the slot is free.
+  ## cleared.
-  let context = sales.context
+  let queue = sales.context.slotQueue
-  let market = context.market
-  let queue = context.slotQueue
 
-  logScope:
+  queue.clearSeenFlags()
-    topics = "marketplace sales onAvailabilityAdded callback"
+  if queue.paused:
-
+    trace "unpausing queue after new availability added"
-  trace "availability added, querying past storage requests to add to queue"
+    queue.unpause()
-
-  try:
-    let events = await market.queryPastStorageRequests(256)
-
-    if events.len == 0:
-      trace "no storage request events found in recent past"
-      return
-
-    let requests = events.map(event =>
-      SlotQueueItem.init(event.requestId, event.ask, event.expiry)
-    )
-
-    trace "found past storage requested events to add to queue",
-      events = events.len
-
-    for slots in requests:
-      for slot in slots:
-        if err =? queue.push(slot).errorOption:
-          # continue on error
-          if err of QueueNotRunningError:
-            warn "cannot push items to queue, queue is not running"
-          elif err of NoMatchingAvailabilityError:
-            info "slot in queue had no matching availabilities, ignoring"
-          elif err of SlotsOutOfRangeError:
-            warn "Too many slots, cannot add to queue"
-          elif err of SlotQueueItemExistsError:
-            trace "item already exists, ignoring"
-            discard
-          else: raise err
-  except CancelledError as error:
-    raise error
-  except CatchableError as e:
-    warn "Error adding request to SlotQueue", error = e.msg
-    discard
 
 proc onStorageRequested(sales: Sales,
                         requestId: RequestId,
@@ -321,9 +303,7 @@ proc onStorageRequested(sales: Sales,
   for item in items:
     # continue on failure
     if err =? slotQueue.push(item).errorOption:
-      if err of NoMatchingAvailabilityError:
+      if err of SlotQueueItemExistsError:
-        info "slot in queue had no matching availabilities, ignoring"
-      elif err of SlotQueueItemExistsError:
         error "Failed to push item to queue becaue it already exists"
       elif err of QueueNotRunningError:
         warn "Failed to push item to queue becaue queue is not running"
@@ -364,9 +344,7 @@ proc onSlotFreed(sales: Sales,
   addSlotToQueue()
     .track(sales)
     .catch(proc(err: ref CatchableError) =
-      if err of NoMatchingAvailabilityError:
+      if err of SlotQueueItemExistsError:
-        info "slot in queue had no matching availabilities, ignoring"
-      elif err of SlotQueueItemExistsError:
         error "Failed to push item to queue becaue it already exists"
       elif err of QueueNotRunningError:
         warn "Failed to push item to queue becaue queue is not running"
@@ -489,6 +467,7 @@ proc startSlotQueue(sales: Sales) {.async.} =
 
   slotQueue.onProcessSlot =
     proc(item: SlotQueueItem, done: Future[void]) {.async.} =
+      trace "processing slot queue item", reqId = item.requestId, slotIdx = item.slotIndex
       sales.processSlot(item, done)
 
   asyncSpawn slotQueue.start()

codex/sales/reservations.nim

@@ -28,6 +28,8 @@
 import pkg/upraises
 push: {.upraises: [].}
 
+import std/sequtils
+import std/sugar
 import std/typetraits
 import std/sequtils
 import pkg/chronos
@@ -37,6 +39,7 @@ import pkg/questionable
 import pkg/questionable/results
 import pkg/stint
 import pkg/stew/byteutils
+import ../codextypes
 import ../logutils
 import ../clock
 import ../stores
@@ -90,6 +93,8 @@ const
   SalesKey = (CodexMetaKey / "sales").tryGet # TODO: move to sales module
   ReservationsKey = (SalesKey / "reservations").tryGet
 
+proc all*(self: Reservations, T: type SomeStorableObject): Future[?!seq[T]] {.async.}
+
 proc new*(T: type Reservations,
           repo: RepoStore): Reservations =
 
@@ -226,10 +231,26 @@ proc update*(
   without key =? obj.key, error:
     return failure(error)
 
-  let getResult = await self.get(key, Availability)
+  without oldAvailability =? await self.get(key, Availability), err:
-
+    if err of NotExistsError:
-  if getResult.isOk:
+      let res = await self.updateImpl(obj)
-    let oldAvailability = !getResult
+      # inform subscribers that Availability has been added
+      if onAvailabilityAdded =? self.onAvailabilityAdded:
+        # when chronos v4 is implemented, and OnAvailabilityAdded is annotated
+        # with async:(raises:[]), we can remove this try/catch as we know, with
+        # certainty, that nothing will be raised
+        try:
+          await onAvailabilityAdded(obj)
+        except CancelledError as e:
+          raise e
+        except CatchableError as e:
+          # we don't have any insight into types of exceptions that
+          # `onAvailabilityAdded` can raise because it is caller-defined
+          warn "Unknown error during 'onAvailabilityAdded' callback",
+            availabilityId = obj.id, error = e.msg
+      return res
+    else:
+      return failure(err)
 
   # Sizing of the availability changed, we need to adjust the repo reservation accordingly
   if oldAvailability.totalSize != obj.totalSize:
@@ -240,12 +261,27 @@ proc update*(
     elif oldAvailability.totalSize > obj.totalSize: # storage removed
       if reserveErr =? (await self.repo.release((oldAvailability.totalSize - obj.totalSize).truncate(uint))).errorOption:
         return failure(reserveErr.toErr(ReleaseFailedError))
-  else:
-    let err = getResult.error()
-    if not (err of NotExistsError):
-      return failure(err)
 
-  return await self.updateImpl(obj)
+  let res = await self.updateImpl(obj)
+
+  if oldAvailability.freeSize < obj.freeSize: # availability added
+    # inform subscribers that Availability has been modified (with increased
+    # size)
+    if onAvailabilityAdded =? self.onAvailabilityAdded:
+      # when chronos v4 is implemented, and OnAvailabilityAdded is annotated
+      # with async:(raises:[]), we can remove this try/catch as we know, with
+      # certainty, that nothing will be raised
+      try:
+        await onAvailabilityAdded(obj)
+      except CancelledError as e:
+        raise e
+      except CatchableError as e:
+        # we don't have any insight into types of exceptions that
+        # `onAvailabilityAdded` can raise because it is caller-defined
+        warn "Unknown error during 'onAvailabilityAdded' callback",
+          availabilityId = obj.id, error = e.msg
+
+  return res
 
 proc delete(
   self: Reservations,
@@ -300,6 +336,9 @@ proc deleteReservation*(
 
   return success()
 
+# TODO: add support for deleting availabilities
+# To delete, must not have any active sales.
+
 proc createAvailability*(
   self: Reservations,
   size: UInt256,
@@ -327,17 +366,6 @@ proc createAvailability*(
 
     return failure(updateErr)
 
-  if onAvailabilityAdded =? self.onAvailabilityAdded:
-    try:
-      await onAvailabilityAdded(availability)
-    except CancelledError as error:
-      raise error
-    except CatchableError as e:
-      # we don't have any insight into types of errors that `onProcessSlot` can
-      # throw because it is caller-defined
-      warn "Unknown error during 'onAvailabilityAdded' callback",
-        availabilityId = availability.id, error = e.msg
-
   return success(availability)
 
 proc createReservation*(

codex/sales/salesagent.nim

@@ -25,7 +25,7 @@ type
     onCleanUp*: OnCleanUp
     onFilled*: ?OnFilled
 
-  OnCleanUp* = proc (returnBytes = false): Future[void] {.gcsafe, upraises: [].}
+  OnCleanUp* = proc (returnBytes = false, reprocessSlot = false): Future[void] {.gcsafe, upraises: [].}
   OnFilled* = proc(request: StorageRequest,
                    slotIndex: UInt256) {.gcsafe, upraises: [].}
 

codex/sales/slotqueue.nim

@@ -36,6 +36,7 @@ type
     reward: UInt256
     collateral: UInt256
     expiry: UInt256
+    seen: bool
 
   # don't need to -1 to prevent overflow when adding 1 (to always allow push)
   # because AsyncHeapQueue size is of type `int`, which is larger than `uint16`
@@ -48,12 +49,12 @@ type
     running: bool
     workers: AsyncQueue[SlotQueueWorker]
     trackedFutures: TrackedFutures
+    unpaused: AsyncEvent
 
   SlotQueueError = object of CodexError
   SlotQueueItemExistsError* = object of SlotQueueError
   SlotQueueItemNotExistsError* = object of SlotQueueError
   SlotsOutOfRangeError* = object of SlotQueueError
-  NoMatchingAvailabilityError* = object of SlotQueueError
   QueueNotRunningError* = object of SlotQueueError
 
 # Number of concurrent workers used for processing SlotQueueItems
@@ -84,6 +85,9 @@ proc `<`*(a, b: SlotQueueItem): bool =
     if condition:
       score += 1'u8 shl addition
 
+  scoreA.addIf(a.seen < b.seen, 4)
+  scoreB.addIf(a.seen > b.seen, 4)
+
   scoreA.addIf(a.profitability > b.profitability, 3)
   scoreB.addIf(a.profitability < b.profitability, 3)
 
@@ -117,12 +121,13 @@ proc new*(_: type SlotQueue,
     # temporarily. After push (and sort), the bottom-most item will be deleted
     queue: newAsyncHeapQueue[SlotQueueItem](maxSize.int + 1),
     running: false,
-    trackedFutures: TrackedFutures.new()
+    trackedFutures: TrackedFutures.new(),
+    unpaused: newAsyncEvent()
   )
   # avoid instantiating `workers` in constructor to avoid side effects in
   # `newAsyncQueue` procedure
 
-proc init*(_: type SlotQueueWorker): SlotQueueWorker =
+proc init(_: type SlotQueueWorker): SlotQueueWorker =
   SlotQueueWorker(
     doneProcessing: newFuture[void]("slotqueue.worker.processing")
   )
@@ -131,7 +136,8 @@ proc init*(_: type SlotQueueItem,
           requestId: RequestId,
           slotIndex: uint16,
           ask: StorageAsk,
-          expiry: UInt256): SlotQueueItem =
+          expiry: UInt256,
+          seen = false): SlotQueueItem =
 
   SlotQueueItem(
     requestId: requestId,
@@ -140,7 +146,8 @@ proc init*(_: type SlotQueueItem,
     duration: ask.duration,
     reward: ask.reward,
     collateral: ask.collateral,
-    expiry: expiry
+    expiry: expiry,
+    seen: seen
   )
 
 proc init*(_: type SlotQueueItem,
@@ -184,6 +191,7 @@ proc slotSize*(self: SlotQueueItem): UInt256 = self.slotSize
 proc duration*(self: SlotQueueItem): UInt256 = self.duration
 proc reward*(self: SlotQueueItem): UInt256 = self.reward
 proc collateral*(self: SlotQueueItem): UInt256 = self.collateral
+proc seen*(self: SlotQueueItem): bool = self.seen
 
 proc running*(self: SlotQueue): bool = self.running
 
@@ -191,6 +199,8 @@ proc len*(self: SlotQueue): int = self.queue.len
 
 proc size*(self: SlotQueue): int = self.queue.size - 1
 
+proc paused*(self: SlotQueue): bool = not self.unpaused.isSet
+
 proc `$`*(self: SlotQueue): string = $self.queue
 
 proc `onProcessSlot=`*(self: SlotQueue, onProcessSlot: OnProcessSlot) =
@@ -205,6 +215,14 @@ proc activeWorkers*(self: SlotQueue): int =
 proc contains*(self: SlotQueue, item: SlotQueueItem): bool =
   self.queue.contains(item)
 
+proc pause*(self: SlotQueue) =
+  # set unpaused flag to false -- coroutines will block on unpaused.wait()
+  self.unpaused.clear()
+
+proc unpause*(self: SlotQueue) =
+  # set unpaused flag to true -- unblocks coroutines waiting on unpaused.wait()
+  self.unpaused.fire()
+
 proc populateItem*(self: SlotQueue,
                    requestId: RequestId,
                    slotIndex: uint16): ?SlotQueueItem =
@@ -226,8 +244,12 @@ proc populateItem*(self: SlotQueue,
 
 proc push*(self: SlotQueue, item: SlotQueueItem): ?!void =
 
-  trace "pushing item to queue",
+  logScope:
-    requestId = item.requestId, slotIndex = item.slotIndex
+    requestId = item.requestId
+    slotIndex = item.slotIndex
+    seen = item.seen
+
+  trace "pushing item to queue"
 
   if not self.running:
     let err = newException(QueueNotRunningError, "queue not running")
@@ -245,6 +267,13 @@ proc push*(self: SlotQueue, item: SlotQueueItem): ?!void =
     self.queue.del(self.queue.size - 1)
 
   doAssert self.queue.len <= self.queue.size - 1
+
+  # when slots are pushed to the queue, the queue should be unpaused if it was
+  # paused
+  if self.paused and not item.seen:
+    trace "unpausing queue after new slot pushed"
+    self.unpause()
+
   return success()
 
 proc push*(self: SlotQueue, items: seq[SlotQueueItem]): ?!void =
@@ -295,6 +324,7 @@ proc addWorker(self: SlotQueue): ?!void =
 
   let worker = SlotQueueWorker.init()
   try:
+    discard worker.doneProcessing.track(self)
     self.workers.addLastNoWait(worker)
   except AsyncQueueFullError:
     return failure("failed to add worker, worker queue full")
@@ -314,6 +344,7 @@ proc dispatch(self: SlotQueue,
 
   if onProcessSlot =? self.onProcessSlot:
     try:
+      discard worker.doneProcessing.track(self)
       await onProcessSlot(item, worker.doneProcessing)
       await worker.doneProcessing
 
@@ -332,6 +363,23 @@ proc dispatch(self: SlotQueue,
       # throw because it is caller-defined
       warn "Unknown error processing slot in worker", error = e.msg
 
+proc clearSeenFlags*(self: SlotQueue) =
+  # Enumerate all items in the queue, overwriting each item with `seen = false`.
+  # To avoid issues with new queue items being pushed to the queue while all
+  # items are being iterated (eg if a new storage request comes in and pushes
+  # new slots to the queue), this routine must remain synchronous.
+
+  if self.queue.empty:
+    return
+
+  for item in self.queue.mitems:
+    item.seen = false # does not maintain the heap invariant
+
+  # force heap reshuffling to maintain the heap invariant
+  doAssert self.queue.update(self.queue[0]), "slot queue failed to reshuffle"
+
+  trace "all 'seen' flags cleared"
+
 proc start*(self: SlotQueue) {.async.} =
   if self.running:
     return
@@ -351,21 +399,47 @@ proc start*(self: SlotQueue) {.async.} =
 
   while self.running:
     try:
+      if self.paused:
+        trace "Queue is paused, waiting for new slots or availabilities to be modified/added"
+
+      # block until unpaused is true/fired, ie wait for queue to be unpaused
+      await self.unpaused.wait()
+
       let worker = await self.workers.popFirst().track(self) # if workers saturated, wait here for new workers
       let item = await self.queue.pop().track(self) # if queue empty, wait here for new items
 
+      logScope:
+        reqId = item.requestId
+        slotIdx = item.slotIndex
+        seen = item.seen
+
       if not self.running: # may have changed after waiting for pop
         trace "not running, exiting"
         break
 
+      # If, upon processing a slot, the slot item already has a `seen` flag set,
+      # the queue should be paused.
+      if item.seen:
+        trace "processing already seen item, pausing queue",
+          reqId = item.requestId, slotIdx = item.slotIndex
+        self.pause()
+        # put item back in queue so that if other items are pushed while paused,
+        # it will be sorted accordingly. Otherwise, this item would be processed
+        # immediately (with priority over other items) once unpaused
+        trace "readding seen item back into the queue"
+        discard self.push(item) # on error, drop the item and continue
+        worker.doneProcessing.complete()
+        await sleepAsync(1.millis) # poll
+        continue
+
+      trace "processing item"
+
       self.dispatch(worker, item)
         .track(self)
         .catch(proc (e: ref CatchableError) =
           error "Unknown error dispatching worker", error = e.msg
         )
 
-      discard worker.doneProcessing.track(self)
-
       await sleepAsync(1.millis) # poll
     except CancelledError:
       trace "slot queue cancelled"

codex/sales/states/cancelled.nim

@@ -28,6 +28,6 @@ method run*(state: SaleCancelled, machine: Machine): Future[?State] {.async.} =
     onClear(request, data.slotIndex)
 
   if onCleanUp =? agent.onCleanUp:
-    await onCleanUp(returnBytes = true)
+    await onCleanUp(returnBytes = true, reprocessSlot = false)
 
   warn "Sale cancelled due to timeout",  requestId = data.requestId, slotIndex = data.slotIndex

codex/sales/states/downloading.nim

@@ -69,7 +69,7 @@ method run*(state: SaleDownloading, machine: Machine): Future[?State] {.async.}
   if err =? (await onStore(request,
                            data.slotIndex,
                            onBlocks)).errorOption:
-    return some State(SaleErrored(error: err))
+    return some State(SaleErrored(error: err, reprocessSlot: false))
 
   trace "Download complete"
   return some State(SaleInitialProving())

codex/sales/states/errored.nim

@@ -12,6 +12,7 @@ logScope:
 
 type SaleErrored* = ref object of SaleState
   error*: ref CatchableError
+  reprocessSlot*: bool
 
 method `$`*(state: SaleErrored): string = "SaleErrored"
 
@@ -30,5 +31,5 @@ method run*(state: SaleErrored, machine: Machine): Future[?State] {.async.} =
     onClear(request, data.slotIndex)
 
   if onCleanUp =? agent.onCleanUp:
-    await onCleanUp(returnBytes = true)
+    await onCleanUp(returnBytes = true, reprocessSlot = state.reprocessSlot)
 

codex/sales/states/ignored.nim

@@ -17,4 +17,7 @@ method run*(state: SaleIgnored, machine: Machine): Future[?State] {.async.} =
   let agent = SalesAgent(machine)
 
   if onCleanUp =? agent.onCleanUp:
-    await onCleanUp()
+    # Ignored slots mean there was no availability. In order to prevent small
+    # availabilities from draining the queue, mark this slot as seen and re-add
+    # back into the queue.
+    await onCleanUp(reprocessSlot = true)

tests/codex/helpers/mockslotqueueitem.nim

@@ -0,0 +1,26 @@
+import pkg/codex/contracts/requests
+import pkg/codex/sales/slotqueue
+
+type MockSlotQueueItem* = object
+  requestId*: RequestId
+  slotIndex*: uint16
+  slotSize*: UInt256
+  duration*: UInt256
+  reward*: UInt256
+  collateral*: UInt256
+  expiry*: UInt256
+  seen*: bool
+
+proc toSlotQueueItem*(item: MockSlotQueueItem): SlotQueueItem =
+  SlotQueueItem.init(
+    requestId = item.requestId,
+    slotIndex = item.slotIndex,
+    ask = StorageAsk(
+            slotSize: item.slotSize,
+            duration: item.duration,
+            reward: item.reward,
+            collateral: item.collateral
+          ),
+    expiry = item.expiry,
+    seen = item.seen
+  )

tests/codex/sales/states/testcancelled.nim

@@ -0,0 +1,45 @@
+import pkg/questionable
+import pkg/chronos
+import pkg/codex/contracts/requests
+import pkg/codex/sales/states/cancelled
+import pkg/codex/sales/salesagent
+import pkg/codex/sales/salescontext
+import pkg/codex/market
+
+import ../../../asynctest
+import ../../examples
+import ../../helpers
+import ../../helpers/mockmarket
+import ../../helpers/mockclock
+
+asyncchecksuite "sales state 'cancelled'":
+  let request = StorageRequest.example
+  let slotIndex = (request.ask.slots div 2).u256
+  let market = MockMarket.new()
+  let clock = MockClock.new()
+
+  var state: SaleCancelled
+  var agent: SalesAgent
+  var returnBytesWas = false
+  var reprocessSlotWas = false
+
+  setup:
+    let onCleanUp = proc (returnBytes = false, reprocessSlot = false) {.async.} =
+                      returnBytesWas = returnBytes
+                      reprocessSlotWas = reprocessSlot
+
+    let context = SalesContext(
+      market: market,
+      clock: clock
+    )
+    agent = newSalesAgent(context,
+                          request.id,
+                          slotIndex,
+                          request.some)
+    agent.onCleanUp = onCleanUp
+    state = SaleCancelled.new()
+
+  test "calls onCleanUp with returnBytes = false and reprocessSlot = true":
+    discard await state.run(agent)
+    check eventually returnBytesWas == true
+    check eventually reprocessSlotWas == false

tests/codex/sales/states/testdownloading.nim

@@ -1,8 +1,9 @@
 import std/unittest
 import pkg/questionable
 import pkg/codex/contracts/requests
-import pkg/codex/sales/states/downloading
 import pkg/codex/sales/states/cancelled
+import pkg/codex/sales/states/downloading
+import pkg/codex/sales/states/errored
 import pkg/codex/sales/states/failed
 import pkg/codex/sales/states/filled
 import ../../examples

tests/codex/sales/states/testerrored.nim

@@ -0,0 +1,49 @@
+import pkg/questionable
+import pkg/chronos
+import pkg/codex/contracts/requests
+import pkg/codex/sales/states/errored
+import pkg/codex/sales/salesagent
+import pkg/codex/sales/salescontext
+import pkg/codex/market
+
+import ../../../asynctest
+import ../../examples
+import ../../helpers
+import ../../helpers/mockmarket
+import ../../helpers/mockclock
+
+asyncchecksuite "sales state 'errored'":
+  let request = StorageRequest.example
+  let slotIndex = (request.ask.slots div 2).u256
+  let market = MockMarket.new()
+  let clock = MockClock.new()
+
+  var state: SaleErrored
+  var agent: SalesAgent
+  var returnBytesWas = false
+  var reprocessSlotWas = false
+
+  setup:
+    let onCleanUp = proc (returnBytes = false, reprocessSlot = false) {.async.} =
+                      returnBytesWas = returnBytes
+                      reprocessSlotWas = reprocessSlot
+
+    let context = SalesContext(
+      market: market,
+      clock: clock
+    )
+    agent = newSalesAgent(context,
+                          request.id,
+                          slotIndex,
+                          request.some)
+    agent.onCleanUp = onCleanUp
+    state = SaleErrored(error: newException(ValueError, "oh no!"))
+
+  test "calls onCleanUp with returnBytes = false and reprocessSlot = true":
+    state = SaleErrored(
+      error: newException(ValueError, "oh no!"),
+      reprocessSlot: true
+    )
+    discard await state.run(agent)
+    check eventually returnBytesWas == true
+    check eventually reprocessSlotWas == true

tests/codex/sales/states/testignored.nim

@@ -0,0 +1,45 @@
+import pkg/questionable
+import pkg/chronos
+import pkg/codex/contracts/requests
+import pkg/codex/sales/states/ignored
+import pkg/codex/sales/salesagent
+import pkg/codex/sales/salescontext
+import pkg/codex/market
+
+import ../../../asynctest
+import ../../examples
+import ../../helpers
+import ../../helpers/mockmarket
+import ../../helpers/mockclock
+
+asyncchecksuite "sales state 'ignored'":
+  let request = StorageRequest.example
+  let slotIndex = (request.ask.slots div 2).u256
+  let market = MockMarket.new()
+  let clock = MockClock.new()
+
+  var state: SaleIgnored
+  var agent: SalesAgent
+  var returnBytesWas = false
+  var reprocessSlotWas = false
+
+  setup:
+    let onCleanUp = proc (returnBytes = false, reprocessSlot = false) {.async.} =
+                      returnBytesWas = returnBytes
+                      reprocessSlotWas = reprocessSlot
+
+    let context = SalesContext(
+      market: market,
+      clock: clock
+    )
+    agent = newSalesAgent(context,
+                          request.id,
+                          slotIndex,
+                          request.some)
+    agent.onCleanUp = onCleanUp
+    state = SaleIgnored.new()
+
+  test "calls onCleanUp with returnBytes = false and reprocessSlot = true":
+    discard await state.run(agent)
+    check eventually returnBytesWas == false
+    check eventually reprocessSlotWas == true

tests/codex/sales/testreservations.nim

@@ -258,7 +258,7 @@ asyncchecksuite "Reservations module":
     check updated.isErr
     check updated.error of NotExistsError
 
-  test "onAvailabilityAdded called when availability is reserved":
+  test "onAvailabilityAdded called when availability is created":
     var added: Availability
     reservations.onAvailabilityAdded = proc(a: Availability) {.async.} =
       added = a
@@ -267,6 +267,26 @@ asyncchecksuite "Reservations module":
 
     check added == availability
 
+  test "onAvailabilityAdded called when availability size is increased":
+    var availability = createAvailability()
+    var added: Availability
+    reservations.onAvailabilityAdded = proc(a: Availability) {.async.} =
+      added = a
+    availability.freeSize += 1.u256
+    discard await reservations.update(availability)
+
+    check added == availability
+
+  test "onAvailabilityAdded is not called when availability size is decreased":
+    var availability = createAvailability()
+    var called = false
+    reservations.onAvailabilityAdded = proc(a: Availability) {.async.} =
+      called = true
+    availability.freeSize -= 1.u256
+    discard await reservations.update(availability)
+
+    check not called
+
   test "availabilities can be found":
     let availability = createAvailability()
 

tests/codex/sales/testsales.nim

@@ -272,24 +272,41 @@ asyncchecksuite "Sales":
     let expected = SlotQueueItem.init(request, 2.uint16)
     check eventually itemsProcessed.contains(expected)
 
-  test "adds past requests to queue once availability added":
+  test "items in queue are readded (and marked seen) once ignored":
-    var itemsProcessed: seq[SlotQueueItem] = @[]
-
-    # ignore all
-    queue.onProcessSlot = proc(item: SlotQueueItem, done: Future[void]) {.async.} =
-      done.complete()
-
     await market.requestStorage(request)
-    await sleepAsync(10.millis)
+    let items = SlotQueueItem.init(request)
+    await sleepAsync(10.millis) # queue starts paused, allow items to be added to the queue
+    check eventually queue.paused
+    # The first processed item will be will have been re-pushed with `seen =
+    # true`. Then, once this item is processed by the queue, its 'seen' flag
+    # will be checked, at which point the queue will be paused. This test could
+    # check item existence in the queue, but that would require inspecting
+    # onProcessSlot to see which item was first, and overridding onProcessSlot
+    # will prevent the queue working as expected in the Sales module.
+    check eventually queue.len == 4
 
-    # check how many slots were processed by the queue
+    for item in items:
-    queue.onProcessSlot = proc(item: SlotQueueItem, done: Future[void]) {.async.} =
+      check queue.contains(item)
-      itemsProcessed.add item
-      done.complete()
 
-    # now add matching availability
+    for i in 0..<queue.len:
-    createAvailability()
+      check queue[i].seen
-    check eventually itemsProcessed.len == request.ask.slots.int
+
+  test "queue is paused once availability is insufficient to service slots in queue":
+    createAvailability() # enough to fill a single slot
+    await market.requestStorage(request)
+    let items = SlotQueueItem.init(request)
+    await sleepAsync(10.millis) # queue starts paused, allow items to be added to the queue
+    check eventually queue.paused
+    # The first processed item/slot will be filled (eventually). Subsequent
+    # items will be processed and eventually re-pushed with `seen = true`. Once
+    # a "seen" item is processed by the queue, the queue is paused. In the
+    # meantime, the other items that are process, marked as seen, and re-added
+    # to the queue may be processed simultaneously as the queue pausing.
+    # Therefore, there should eventually be 3 items remaining in the queue, all
+    # seen.
+    check eventually queue.len == 3
+    for i in 0..<queue.len:
+      check queue[i].seen
 
   test "availability size is reduced by request slot size when fully downloaded":
     sales.onStore = proc(request: StorageRequest,
@@ -495,6 +512,10 @@ asyncchecksuite "Sales":
 
   test "verifies that request is indeed expired from onchain before firing onCancelled":
     let expiry = getTime().toUnix() + 10
+    # ensure only one slot, otherwise once bytes are returned to the
+    # availability, the queue will be unpaused and availability will be consumed
+    # by other slots
+    request.ask.slots = 1.uint64
     market.requestExpiry[request.id] = expiry
 
     let origSize = availability.freeSize

tests/codex/sales/testslotqueue.nim

@@ -10,6 +10,7 @@ import pkg/codex/sales/slotqueue
 import ../../asynctest
 import ../helpers
 import ../helpers/mockmarket
+import ../helpers/mockslotqueueitem
 import ../examples
 
 suite "Slot queue start/stop":
@@ -118,7 +119,6 @@ suite "Slot queue":
     queue = SlotQueue.new(maxWorkers, maxSize.uint16)
     queue.onProcessSlot = proc(item: SlotQueueItem, done: Future[void]) {.async.} =
       await sleepAsync(processSlotDelay)
-      trace "processing item", requestId = item.requestId, slotIndex = item.slotIndex
       onProcessSlotCalled = true
       onProcessSlotCalledWith.add (item.requestId, item.slotIndex)
       done.complete()
@@ -162,6 +162,131 @@ suite "Slot queue":
     check itemB < itemA # B higher priority than A
     check itemA > itemB
 
+  test "correct prioritizes SlotQueueItems based on 'seen'":
+    let request = StorageRequest.example
+    let itemA = MockSlotQueueItem(
+      requestId: request.id,
+      slotIndex: 0,
+      slotSize: 1.u256,
+      duration: 1.u256,
+      reward: 2.u256, # profitability is higher (good)
+      collateral: 1.u256,
+      expiry: 1.u256,
+      seen: true # seen (bad), more weight than profitability
+    )
+    let itemB = MockSlotQueueItem(
+      requestId: request.id,
+      slotIndex: 0,
+      slotSize: 1.u256,
+      duration: 1.u256,
+      reward: 1.u256, # profitability is lower (bad)
+      collateral: 1.u256,
+      expiry: 1.u256,
+      seen: false # not seen (good)
+    )
+    check itemB.toSlotQueueItem < itemA.toSlotQueueItem # B higher priority than A
+    check itemA.toSlotQueueItem > itemB.toSlotQueueItem
+
+  test "correct prioritizes SlotQueueItems based on profitability":
+    let request = StorageRequest.example
+    let itemA = MockSlotQueueItem(
+      requestId: request.id,
+      slotIndex: 0,
+      slotSize: 1.u256,
+      duration: 1.u256,
+      reward: 1.u256, # reward is lower (bad)
+      collateral: 1.u256, # collateral is lower (good)
+      expiry: 1.u256,
+      seen: false
+    )
+    let itemB = MockSlotQueueItem(
+      requestId: request.id,
+      slotIndex: 0,
+      slotSize: 1.u256,
+      duration: 1.u256,
+      reward: 2.u256, # reward is higher (good), more weight than collateral
+      collateral: 2.u256, # collateral is higher (bad)
+      expiry: 1.u256,
+      seen: false
+    )
+
+    check itemB.toSlotQueueItem < itemA.toSlotQueueItem # < indicates higher priority
+
+  test "correct prioritizes SlotQueueItems based on collateral":
+    let request = StorageRequest.example
+    let itemA = MockSlotQueueItem(
+      requestId: request.id,
+      slotIndex: 0,
+      slotSize: 1.u256,
+      duration: 1.u256,
+      reward: 1.u256,
+      collateral: 2.u256, # collateral is higher (bad)
+      expiry: 2.u256, # expiry is longer (good)
+      seen: false
+    )
+    let itemB = MockSlotQueueItem(
+      requestId: request.id,
+      slotIndex: 0,
+      slotSize: 1.u256,
+      duration: 1.u256,
+      reward: 1.u256,
+      collateral: 1.u256, # collateral is lower (good), more weight than expiry
+      expiry: 1.u256, # expiry is shorter (bad)
+      seen: false
+    )
+
+    check itemB.toSlotQueueItem < itemA.toSlotQueueItem # < indicates higher priority
+
+  test "correct prioritizes SlotQueueItems based on expiry":
+    let request = StorageRequest.example
+    let itemA = MockSlotQueueItem(
+      requestId: request.id,
+      slotIndex: 0,
+      slotSize: 1.u256, # slotSize is smaller (good)
+      duration: 1.u256,
+      reward: 1.u256,
+      collateral: 1.u256,
+      expiry: 1.u256, # expiry is shorter (bad)
+      seen: false
+    )
+    let itemB = MockSlotQueueItem(
+      requestId: request.id,
+      slotIndex: 0,
+      slotSize: 2.u256, # slotSize is larger (bad)
+      duration: 1.u256,
+      reward: 1.u256,
+      collateral: 1.u256,
+      expiry: 2.u256, # expiry is longer (good), more weight than slotSize
+      seen: false
+    )
+
+    check itemB.toSlotQueueItem < itemA.toSlotQueueItem # < indicates higher priority
+
+  test "correct prioritizes SlotQueueItems based on slotSize":
+    let request = StorageRequest.example
+    let itemA = MockSlotQueueItem(
+      requestId: request.id,
+      slotIndex: 0,
+      slotSize: 2.u256, # slotSize is larger (bad)
+      duration: 1.u256,
+      reward: 1.u256,
+      collateral: 1.u256,
+      expiry: 1.u256, # expiry is shorter (bad)
+      seen: false
+    )
+    let itemB = MockSlotQueueItem(
+      requestId: request.id,
+      slotIndex: 0,
+      slotSize: 1.u256, # slotSize is smaller (good)
+      duration: 1.u256,
+      reward: 1.u256,
+      collateral: 1.u256,
+      expiry: 1.u256,
+      seen: false
+    )
+
+    check itemB.toSlotQueueItem < itemA.toSlotQueueItem # < indicates higher priority
+
   test "expands available all possible slot indices on init":
     let request = StorageRequest.example
     let items = SlotQueueItem.init(request)
@@ -391,3 +516,71 @@ suite "Slot queue":
         (item3.requestId, item3.slotIndex),
       ]
     )
+
+  test "processing a 'seen' item pauses the queue":
+    newSlotQueue(maxSize = 4, maxWorkers = 4)
+    let request = StorageRequest.example
+    let item = SlotQueueItem.init(request.id, 0'u16,
+                                  request.ask,
+                                  request.expiry,
+                                  seen = true)
+    queue.push(item)
+    check eventually queue.paused
+    check onProcessSlotCalledWith.len == 0
+
+  test "pushing items to queue unpauses queue":
+    newSlotQueue(maxSize = 4, maxWorkers = 4)
+    queue.pause
+
+    let request = StorageRequest.example
+    var items = SlotQueueItem.init(request)
+    queue.push(items)
+    # check all items processed
+    check eventually queue.len == 0
+
+  test "pushing seen item does not unpause queue":
+    newSlotQueue(maxSize = 4, maxWorkers = 4)
+    let request = StorageRequest.example
+    let item0 = SlotQueueItem.init(request.id, 0'u16,
+                                  request.ask,
+                                  request.expiry,
+                                  seen = true)
+    check queue.paused
+    queue.push(item0)
+    check queue.paused
+
+  test "paused queue waits for unpause before continuing processing":
+    newSlotQueue(maxSize = 4, maxWorkers = 4)
+    let request = StorageRequest.example
+    let item = SlotQueueItem.init(request.id, 1'u16,
+                                  request.ask,
+                                  request.expiry,
+                                  seen = false)
+    check queue.paused
+    # push causes unpause
+    queue.push(item)
+    # check all items processed
+    check eventually onProcessSlotCalledWith == @[
+      (item.requestId, item.slotIndex),
+    ]
+    check eventually queue.len == 0
+
+  test "item 'seen' flags can be cleared":
+    newSlotQueue(maxSize = 4, maxWorkers = 1)
+    let request = StorageRequest.example
+    let item0 = SlotQueueItem.init(request.id, 0'u16,
+                                  request.ask,
+                                  request.expiry,
+                                  seen = true)
+    let item1 = SlotQueueItem.init(request.id, 1'u16,
+                                  request.ask,
+                                  request.expiry,
+                                  seen = true)
+    queue.push(item0)
+    queue.push(item1)
+    check queue[0].seen
+    check queue[1].seen
+
+    queue.clearSeenFlags()
+    check queue[0].seen == false
+    check queue[1].seen == false