cerc-io/lotus
16
0
Fork 0
Code Issues Pull Requests Packages Projects Releases 27 Wiki Activity

Merge pull request #9413 from filecoin-project/sbansal/single-partition-post-for-recovery-sectors

Browse Source 
feat: wdpost: Add ability to only have single partition per msg for partitions with…
This commit is contained in:
Shrenuj Bansal 2022-10-05 10:56:41 +00:00 committed by GitHub
commit e74838f024
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
6 changed files with 166 additions and 51 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
documentation/en/default-lotus-miner-config.toml
View File

@ -380,11 +380,9 @@
# A single partition may contain up to 2349 32GiB sectors, or 2300 64GiB sectors.
#
# The maximum number of sectors which can be proven in a single PoSt message is 25000 in network version 16, which
# means that a single message can prove at most 10 partinions
# means that a single message can prove at most 10 partitions
#
# In some cases when submitting PoSt messages which are recovering sectors, the default network limit may still be
# too high to fit in the block gas limit; In those cases it may be necessary to set this value to something lower
# than 10; Note that setting this value lower may result in less efficient gas use - more messages will be sent,
# Note that setting this value lower may result in less efficient gas use - more messages will be sent,
# to prove each deadline, resulting in more total gas use (but each message will have lower gas limit)
#
# Setting this value above the network limit has no effect
@ -403,6 +401,19 @@
# env var: LOTUS_PROVING_MAXPARTITIONSPERRECOVERYMESSAGE
#MaxPartitionsPerRecoveryMessage = 0
# Enable single partition per PoSt Message for partitions containing recovery sectors
#
# In cases when submitting PoSt messages which contain recovering sectors, the default network limit may still be
# too high to fit in the block gas limit. In those cases, it becomes useful to only house the single partition
# with recovering sectors in the post message
#
# Note that setting this value lower may result in less efficient gas use - more messages will be sent,
# to prove each deadline, resulting in more total gas use (but each message will have lower gas limit)
#
# type: bool
# env var: LOTUS_PROVING_SINGLERECOVERINGPARTITIONPERPOSTMESSAGE
#SingleRecoveringPartitionPerPostMessage = false
[Sealing]
# Upper bound on how many sectors can be waiting for more deals to be packed in it before it begins sealing at any given time.

19
node/config/doc_gen.go
View File

@ -698,11 +698,9 @@ After changing this option, confirm that the new value works in your setup by in
A single partition may contain up to 2349 32GiB sectors, or 2300 64GiB sectors.
The maximum number of sectors which can be proven in a single PoSt message is 25000 in network version 16, which
means that a single message can prove at most 10 partinions
means that a single message can prove at most 10 partitions
In some cases when submitting PoSt messages which are recovering sectors, the default network limit may still be
too high to fit in the block gas limit; In those cases it may be necessary to set this value to something lower
than 10; Note that setting this value lower may result in less efficient gas use - more messages will be sent,
Note that setting this value lower may result in less efficient gas use - more messages will be sent,
to prove each deadline, resulting in more total gas use (but each message will have lower gas limit)
Setting this value above the network limit has no effect`,
@ -717,6 +715,19 @@ In those cases it may be necessary to set this value to something low (eg 1);
Note that setting this value lower may result in less efficient gas use - more messages will be sent than needed,
resulting in more total gas use (but each message will have lower gas limit)`,
},
{
Name: "SingleRecoveringPartitionPerPostMessage",
Type: "bool",
Comment: `Enable single partition per PoSt Message for partitions containing recovery sectors
In cases when submitting PoSt messages which contain recovering sectors, the default network limit may still be
too high to fit in the block gas limit. In those cases, it becomes useful to only house the single partition
with recovering sectors in the post message
Note that setting this value lower may result in less efficient gas use - more messages will be sent,
to prove each deadline, resulting in more total gas use (but each message will have lower gas limit)`,
},
},
"Pubsub": []DocField{
{

16
node/config/types.go
View File

@ -276,11 +276,9 @@ type ProvingConfig struct {
// A single partition may contain up to 2349 32GiB sectors, or 2300 64GiB sectors.
//
// The maximum number of sectors which can be proven in a single PoSt message is 25000 in network version 16, which
// means that a single message can prove at most 10 partinions
// means that a single message can prove at most 10 partitions
//
// In some cases when submitting PoSt messages which are recovering sectors, the default network limit may still be
// too high to fit in the block gas limit; In those cases it may be necessary to set this value to something lower
// than 10; Note that setting this value lower may result in less efficient gas use - more messages will be sent,
// Note that setting this value lower may result in less efficient gas use - more messages will be sent,
// to prove each deadline, resulting in more total gas use (but each message will have lower gas limit)
//
// Setting this value above the network limit has no effect
@ -294,6 +292,16 @@ type ProvingConfig struct {
// Note that setting this value lower may result in less efficient gas use - more messages will be sent than needed,
// resulting in more total gas use (but each message will have lower gas limit)
MaxPartitionsPerRecoveryMessage int
// Enable single partition per PoSt Message for partitions containing recovery sectors
//
// In cases when submitting PoSt messages which contain recovering sectors, the default network limit may still be
// too high to fit in the block gas limit. In those cases, it becomes useful to only house the single partition
// with recovering sectors in the post message
//
// Note that setting this value lower may result in less efficient gas use - more messages will be sent,
// to prove each deadline, resulting in more total gas use (but each message will have lower gas limit)
SingleRecoveringPartitionPerPostMessage bool
}
type SealingConfig struct {

44
storage/wdpost/wdpost_run.go
View File

@ -286,7 +286,7 @@ func (s *WindowPoStScheduler) runPoStCycle(ctx context.Context, manual bool, di
// Split partitions into batches, so as not to exceed the number of sectors
// allowed in a single message
partitionBatches, err := s.batchPartitions(partitions, nv)
partitionBatches, err := s.BatchPartitions(partitions, nv)
if err != nil {
return nil, err
}
@ -492,7 +492,9 @@ func (s *WindowPoStScheduler) runPoStCycle(ctx context.Context, manual bool, di
return posts, nil
}
func (s *WindowPoStScheduler) batchPartitions(partitions []api.Partition, nv network.Version) ([][]api.Partition, error) {
// Note: Partition order within batches must match original partition order in order
// for code following the user code to work
func (s *WindowPoStScheduler) BatchPartitions(partitions []api.Partition, nv network.Version) ([][]api.Partition, error) {
// We don't want to exceed the number of sectors allowed in a message.
// So given the number of sectors in a partition, work out the number of
// partitions that can be in a message without exceeding sectors per
@ -524,21 +526,33 @@ func (s *WindowPoStScheduler) batchPartitions(partitions []api.Partition, nv net
}
}
// The number of messages will be:
// ceiling(number of partitions / partitions per message)
batchCount := len(partitions) / partitionsPerMsg
if len(partitions)%partitionsPerMsg != 0 {
batchCount++
}
batches := [][]api.Partition{}
// Split the partitions into batches
batches := make([][]api.Partition, 0, batchCount)
for i := 0; i < len(partitions); i += partitionsPerMsg {
end := i + partitionsPerMsg
if end > len(partitions) {
end = len(partitions)
currBatch := []api.Partition{}
for _, partition := range partitions {
recSectors, err := partition.RecoveringSectors.Count()
if err != nil {
return nil, err
}
batches = append(batches, partitions[i:end])
// Only add single partition to a batch if it contains recovery sectors
// and has the below user config set
if s.singleRecoveringPartitionPerPostMessage && recSectors > 0 {
if len(currBatch) > 0 {
batches = append(batches, currBatch)
currBatch = []api.Partition{}
}
batches = append(batches, []api.Partition{partition})
} else {
if len(currBatch) >= partitionsPerMsg {
batches = append(batches, currBatch)
currBatch = []api.Partition{}
}
currBatch = append(currBatch, partition)
}
}
if len(currBatch) > 0 {
batches = append(batches, currBatch)
}
return batches, nil

69
storage/wdpost/wdpost_run_test.go
View File

@ -177,6 +177,26 @@ func (m mockFaultTracker) CheckProvable(ctx context.Context, pp abi.RegisteredPo
return map[abi.SectorID]string{}, nil
}
func generatePartition(sectorCount uint64, recoverySectorCount uint64) api.Partition {
var partition api.Partition
sectors := bitfield.New()
recoverySectors := bitfield.New()
for s := uint64(0); s < sectorCount; s++ {
sectors.Set(s)
}
for s := uint64(0); s < recoverySectorCount; s++ {
recoverySectors.Set(s)
}
partition = api.Partition{
AllSectors: sectors,
FaultySectors: bitfield.New(),
RecoveringSectors: recoverySectors,
LiveSectors: sectors,
ActiveSectors: sectors,
}
return partition
}
// TestWDPostDoPost verifies that doPost will send the correct number of window
// PoST messages for a given number of partitions
func TestWDPostDoPost(t *testing.T) {
@ -368,6 +388,55 @@ func TestWDPostDoPostPartLimitConfig(t *testing.T) {
}
}
// TestBatchPartitionsRecoverySectors tests if the batches with recovery sectors
// contain only single partitions while keeping all the partitions in order
func TestBatchPartitionsRecoverySectors(t *testing.T) {
proofType := abi.RegisteredPoStProof_StackedDrgWindow2KiBV1
postAct := tutils.NewIDAddr(t, 100)
mockStgMinerAPI := newMockStorageMinerAPI()
userPartLimit := 4
scheduler := &WindowPoStScheduler{
api: mockStgMinerAPI,
prover: &mockProver{},
verifier: &mockVerif{},
faultTracker: &mockFaultTracker{},
proofType: proofType,
actor: postAct,
journal: journal.NilJournal(),
addrSel: &ctladdr.AddressSelector{},
maxPartitionsPerPostMessage: userPartLimit,
singleRecoveringPartitionPerPostMessage: true,
}
var partitions []api.Partition
for p := 0; p < 4; p++ {
partitions = append(partitions, generatePartition(100, 0))
}
for p := 0; p < 2; p++ {
partitions = append(partitions, generatePartition(100, 10))
}
for p := 0; p < 6; p++ {
partitions = append(partitions, generatePartition(100, 0))
}
partitions = append(partitions, generatePartition(100, 10))
expectedBatchLens := []int{4, 1, 1, 4, 2, 1}
batches, err := scheduler.BatchPartitions(partitions, network.Version16)
require.NoError(t, err)
require.Equal(t, len(batches), 6)
for i, batch := range batches {
require.Equal(t, len(batch), expectedBatchLens[i])
}
}
// TestWDPostDeclareRecoveriesPartLimitConfig verifies that declareRecoveries will send the correct number of
// DeclareFaultsRecovered messages for a given number of partitions based on user config
func TestWDPostDeclareRecoveriesPartLimitConfig(t *testing.T) {

50
storage/wdpost/wdpost_sched.go
View File

@ -64,18 +64,19 @@ type NodeAPI interface {
// WindowPoStScheduler watches the chain though the changeHandler, which in turn
// turn calls the scheduler when the time arrives to do work.
type WindowPoStScheduler struct {
api NodeAPI
feeCfg config.MinerFeeConfig
addrSel *ctladdr.AddressSelector
prover storiface.ProverPoSt
verifier storiface.Verifier
faultTracker sealer.FaultTracker
proofType abi.RegisteredPoStProof
partitionSectors uint64
disablePreChecks bool
maxPartitionsPerPostMessage int
maxPartitionsPerRecoveryMessage int
ch *changeHandler
api NodeAPI
feeCfg config.MinerFeeConfig
addrSel *ctladdr.AddressSelector
prover storiface.ProverPoSt
verifier storiface.Verifier
faultTracker sealer.FaultTracker
proofType abi.RegisteredPoStProof
partitionSectors uint64
disablePreChecks bool
maxPartitionsPerPostMessage int
maxPartitionsPerRecoveryMessage int
singleRecoveringPartitionPerPostMessage bool
ch *changeHandler
actor address.Address
@ -102,18 +103,19 @@ func NewWindowedPoStScheduler(api NodeAPI,
}
return &WindowPoStScheduler{
api: api,
feeCfg: cfg,
addrSel: as,
prover: sp,
verifier: verif,
faultTracker: ft,
proofType: mi.WindowPoStProofType,
partitionSectors: mi.WindowPoStPartitionSectors,
disablePreChecks: pcfg.DisableWDPoStPreChecks,
maxPartitionsPerPostMessage: pcfg.MaxPartitionsPerPoStMessage,
maxPartitionsPerRecoveryMessage: pcfg.MaxPartitionsPerRecoveryMessage,
actor: actor,
api: api,
feeCfg: cfg,
addrSel: as,
prover: sp,
verifier: verif,
faultTracker: ft,
proofType: mi.WindowPoStProofType,
partitionSectors: mi.WindowPoStPartitionSectors,
disablePreChecks: pcfg.DisableWDPoStPreChecks,
maxPartitionsPerPostMessage: pcfg.MaxPartitionsPerPoStMessage,
maxPartitionsPerRecoveryMessage: pcfg.MaxPartitionsPerRecoveryMessage,
singleRecoveringPartitionPerPostMessage: pcfg.SingleRecoveringPartitionPerPostMessage,
actor: actor,
evtTypes: [...]journal.EventType{
evtTypeWdPoStScheduler: j.RegisterEventType("wdpost", "scheduler"),
evtTypeWdPoStProofs: j.RegisterEventType("wdpost", "proofs_processed"),