From cff77167074966e5512f3b9a8413dd75a5bdae74 Mon Sep 17 00:00:00 2001 From: zenground0 Date: Mon, 10 Jul 2023 15:28:14 -0600 Subject: [PATCH] Tool for measuring partition summaries --- cmd/lotus-shed/cron-count.go | 151 ++++++++++++++++++++++++++++++++++- 1 file changed, 150 insertions(+), 1 deletion(-) diff --git a/cmd/lotus-shed/cron-count.go b/cmd/lotus-shed/cron-count.go index 2b8dc8ebf..33e320dac 100644 --- a/cmd/lotus-shed/cron-count.go +++ b/cmd/lotus-shed/cron-count.go @@ -1,14 +1,18 @@ package main import ( + "encoding/json" "fmt" + "os" "github.com/urfave/cli/v2" "golang.org/x/xerrors" "github.com/filecoin-project/go-address" - + "github.com/filecoin-project/go-state-types/abi" + "github.com/filecoin-project/lotus/api" "github.com/filecoin-project/lotus/build" + "github.com/filecoin-project/lotus/chain/types" lcli "github.com/filecoin-project/lotus/cli" ) @@ -17,6 +21,151 @@ var cronWcCmd = &cli.Command{ Description: "cron stats", Subcommands: []*cli.Command{ minerDeadlineCronCountCmd, + minerDeadlinePartitionMeasurementCmd, + }, +} + +type DeadlineRef struct { + Addr address.Address + Height abi.ChainEpoch +} + +type DeadlineSummary struct { + Partitions []PartitionSummary +} + +type PartitionSummary struct { + Live int + Dead int + Diff PartitionDiff +} + +type PartitionDiff struct { + Faulted int + Recovered int + Killed int +} + +var minerDeadlinePartitionMeasurementCmd = &cli.Command{ + Name: "deadline-summary", + Description: "", + Flags: []cli.Flag{ + &cli.BoolFlag{ + Name: "json", + Usage: "read input as json", + Value: true, + }, + &cli.StringFlag{ + Name: "tipset", + Usage: "specify tipset state to search on (pass comma separated array of cids)", + }, + }, + Action: func(c *cli.Context) error { + // read in values to process + if !c.Bool("json") { + return xerrors.Errorf("unsupported non json input format") + } + var refStream []DeadlineRef + if err := json.NewDecoder(os.Stdin).Decode(&refStream); err != nil { + return xerrors.Errorf("failed to parse input: %w", err) + } + + // go from height and sp addr to deadline partition data + n, acloser, err := lcli.GetFullNodeAPI(c) + if err != nil { + return err + } + defer acloser() + ctx := lcli.ReqContext(c) + + dSummaries := make([]DeadlineSummary, len(refStream)) + for _, ref := range refStream { + // get miner's deadline + tsBefore, err := n.ChainGetTipSetByHeight(ctx, ref.Height, types.EmptyTSK) + if err != nil { + return xerrors.Errorf("failed to get tipset at epoch: %d: %w", ref.Height, err) + } + tsAfter, err := n.ChainGetTipSetByHeight(ctx, ref.Height+1, types.EmptyTSK) + if err != nil { + return xerrors.Errorf("failed to get tipset at epoch %d: %w", ref.Height, err) + } + + dline, err := n.StateMinerProvingDeadline(ctx, ref.Addr, tsBefore.Key()) + if err != nil { + return xerrors.Errorf("failed to read proving deadline: %w", err) + } + + // iterate through all partitions at epoch of processing + var pSummaries []PartitionSummary + psBefore, err := n.StateMinerPartitions(ctx, ref.Addr, dline.Index, tsBefore.Key()) + if err != nil { + return xerrors.Errorf("failed to get partitions: %w", err) + } + psAfter, err := n.StateMinerPartitions(ctx, ref.Addr, dline.Index, tsAfter.Key()) + if err != nil { + return xerrors.Errorf("failed to get partitions: %w", err) + } + if len(psBefore) != len(psAfter) { + return xerrors.Errorf("faield") + } + + type partitionCount struct { + live int + dead int + faulty int + recovering int + } + countPartition := func(p api.Partition) (partitionCount, error) { + liveSectors, err := p.LiveSectors.All(abi.MaxSectorNumber) + if err != nil { + return partitionCount{}, xerrors.Errorf("failed to count live sectors in partition: %w", err) + } + allSectors, err := p.AllSectors.All(abi.MaxSectorNumber) + if err != nil { + return partitionCount{}, xerrors.Errorf("failed to count all sectors in partition: %w", err) + } + faultySectors, err := p.FaultySectors.All(abi.MaxSectorNumber) + if err != nil { + return partitionCount{}, xerrors.Errorf("failed to count faulty sectors in partition: %w", err) + } + recoveringSectors, err := p.RecoveringSectors.All(abi.MaxSectorNumber) + if err != nil { + return partitionCount{}, xerrors.Errorf("failed to count recovering sectors in partition: %w", err) + } + + return partitionCount{ + live: len(liveSectors), + dead: len(allSectors) - len(liveSectors), + faulty: len(faultySectors), + recovering: len(recoveringSectors), + }, nil + } + + for i := 0; i < len(psBefore); i++ { + cntBefore, err := countPartition(psBefore[i]) + if err != nil { + return err + } + cntAfter, err := countPartition(psAfter[i]) + if err != nil { + return err + } + pSummaries = append(pSummaries, PartitionSummary{ + Live: cntBefore.live, + Dead: cntBefore.dead, + Diff: PartitionDiff{ + Faulted: cntAfter.faulty - cntBefore.faulty, + Recovered: cntBefore.recovering - cntAfter.recovering, + Killed: cntAfter.dead - cntBefore.dead, + }, + }) + } + dSummaries = append(dSummaries, DeadlineSummary{Partitions: pSummaries}) + } + + // output partition info + json.NewEncoder(os.Stdout).Encode(dSummaries) + return nil }, }