storj/satellite/rangedloop.go
paul cannon 72189330fd satellite/gracefulexit: revamp graceful exit
Currently, graceful exit is a complicated subsystem that keeps a queue
of all pieces expected to be on a node, and asks the node to transfer
those pieces to other nodes one by one. The complexity of the system
has, unfortunately, led to numerous bugs and unexpected behaviors.

We have decided to remove this entire subsystem and restructure graceful
exit as follows:

* Nodes will signal their intent to exit gracefully
* The satellite will not send any new pieces to gracefully exiting nodes
* Pieces on gracefully exiting nodes will be considered by the repair
  subsystem as "retrievable but unhealthy". They will be repaired off of
  the exiting node as needed.
* After one month (with an appropriately high online score), the node
  will be considered exited, and held amounts for the node will be
  released. The repair worker will continue to fetch pieces from the
  node as long as the node stays online.
* If, at the end of the month, a node's online score is below a certain
  threshold, its graceful exit will fail.

Refs: https://github.com/storj/storj/issues/6042
Change-Id: I52d4e07a4198e9cb2adf5e6cee2cb64d6f9f426b
2023-09-27 08:40:01 +00:00

241 lines
5.9 KiB
Go

// Copyright (C) 2022 Storj Labs, Inc.
// See LICENSE for copying information.
package satellite
import (
"context"
"errors"
"net"
"runtime/pprof"
"github.com/spacemonkeygo/monkit/v3"
"github.com/zeebo/errs"
"go.uber.org/zap"
"golang.org/x/sync/errgroup"
"storj.io/private/debug"
"storj.io/storj/private/lifecycle"
"storj.io/storj/satellite/accounting/nodetally"
"storj.io/storj/satellite/audit"
"storj.io/storj/satellite/gc/piecetracker"
"storj.io/storj/satellite/gracefulexit"
"storj.io/storj/satellite/metabase"
"storj.io/storj/satellite/metabase/rangedloop"
"storj.io/storj/satellite/metrics"
"storj.io/storj/satellite/overlay"
"storj.io/storj/satellite/repair/checker"
)
// RangedLoop is the satellite ranged loop process.
//
// architecture: Peer
type RangedLoop struct {
Log *zap.Logger
DB DB
Servers *lifecycle.Group
Services *lifecycle.Group
Audit struct {
Observer rangedloop.Observer
}
Debug struct {
Listener net.Listener
Server *debug.Server
}
Metrics struct {
Observer rangedloop.Observer
}
Overlay struct {
Service *overlay.Service
}
Repair struct {
Observer *checker.Observer
}
GracefulExit struct {
Observer rangedloop.Observer
}
Accounting struct {
NodeTallyObserver *nodetally.Observer
}
PieceTracker struct {
Observer *piecetracker.Observer
}
RangedLoop struct {
Service *rangedloop.Service
}
}
// NewRangedLoop creates a new satellite ranged loop process.
func NewRangedLoop(log *zap.Logger, db DB, metabaseDB *metabase.DB, config *Config, atomicLogLevel *zap.AtomicLevel) (_ *RangedLoop, err error) {
peer := &RangedLoop{
Log: log,
DB: db,
Servers: lifecycle.NewGroup(log.Named("servers")),
Services: lifecycle.NewGroup(log.Named("services")),
}
{ // setup debug
var err error
if config.Debug.Address != "" {
peer.Debug.Listener, err = net.Listen("tcp", config.Debug.Address)
if err != nil {
withoutStack := errors.New(err.Error())
peer.Log.Debug("failed to start debug endpoints", zap.Error(withoutStack))
}
}
debugConfig := config.Debug
debugConfig.ControlTitle = "RangedLoop"
peer.Debug.Server = debug.NewServerWithAtomicLevel(log.Named("debug"), peer.Debug.Listener, monkit.Default, debugConfig, atomicLogLevel)
peer.Servers.Add(lifecycle.Item{
Name: "debug",
Run: peer.Debug.Server.Run,
Close: peer.Debug.Server.Close,
})
}
{ // setup audit observer
peer.Audit.Observer = audit.NewObserver(log.Named("audit"), db.VerifyQueue(), config.Audit)
}
{ // setup metrics observer
peer.Metrics.Observer = metrics.NewObserver()
}
{ // setup gracefulexit
if config.GracefulExit.Enabled && !config.GracefulExit.TimeBased {
peer.GracefulExit.Observer = gracefulexit.NewObserver(
peer.Log.Named("gracefulexit:observer"),
peer.DB.GracefulExit(),
peer.DB.OverlayCache(),
metabaseDB,
config.GracefulExit,
)
}
}
{ // setup node tally observer
peer.Accounting.NodeTallyObserver = nodetally.NewObserver(
log.Named("accounting:nodetally"),
db.StoragenodeAccounting(),
metabaseDB)
}
{ // setup piece tracker observer
peer.PieceTracker.Observer = piecetracker.NewObserver(
log.Named("piecetracker"),
metabaseDB,
peer.DB.OverlayCache(),
config.PieceTracker,
)
}
{ // setup overlay
placement, err := config.Placement.Parse()
if err != nil {
return nil, err
}
peer.Overlay.Service, err = overlay.NewService(peer.Log.Named("overlay"), peer.DB.OverlayCache(), peer.DB.NodeEvents(), placement.CreateFilters, config.Console.ExternalAddress, config.Console.SatelliteName, config.Overlay)
if err != nil {
return nil, errs.Combine(err, peer.Close())
}
peer.Services.Add(lifecycle.Item{
Name: "overlay",
Run: peer.Overlay.Service.Run,
Close: peer.Overlay.Service.Close,
})
}
{ // setup repair
placement, err := config.Placement.Parse()
if err != nil {
return nil, err
}
if len(config.Checker.RepairExcludedCountryCodes) == 0 {
config.Checker.RepairExcludedCountryCodes = config.Overlay.RepairExcludedCountryCodes
}
peer.Repair.Observer = checker.NewObserver(
peer.Log.Named("repair:checker"),
peer.DB.RepairQueue(),
peer.Overlay.Service,
placement.CreateFilters,
config.Checker,
)
}
{ // setup ranged loop
observers := []rangedloop.Observer{
rangedloop.NewLiveCountObserver(metabaseDB, config.RangedLoop.SuspiciousProcessedRatio, config.RangedLoop.AsOfSystemInterval),
peer.Metrics.Observer,
}
if config.Audit.UseRangedLoop {
observers = append(observers, peer.Audit.Observer)
}
if config.Tally.UseRangedLoop {
observers = append(observers, peer.Accounting.NodeTallyObserver)
}
if peer.GracefulExit.Observer != nil && config.GracefulExit.UseRangedLoop {
observers = append(observers, peer.GracefulExit.Observer)
}
if config.Repairer.UseRangedLoop {
observers = append(observers, peer.Repair.Observer)
}
if config.PieceTracker.UseRangedLoop {
observers = append(observers, peer.PieceTracker.Observer)
}
segments := rangedloop.NewMetabaseRangeSplitter(metabaseDB, config.RangedLoop.AsOfSystemInterval, config.RangedLoop.BatchSize)
peer.RangedLoop.Service = rangedloop.NewService(log.Named("rangedloop"), config.RangedLoop, segments, observers)
peer.Services.Add(lifecycle.Item{
Name: "rangeloop",
Run: peer.RangedLoop.Service.Run,
})
}
return peer, nil
}
// Run runs satellite ranged loop until it's either closed or it errors.
func (peer *RangedLoop) Run(ctx context.Context) (err error) {
defer mon.Task()(&ctx)(&err)
group, ctx := errgroup.WithContext(ctx)
pprof.Do(ctx, pprof.Labels("subsystem", "rangedloop"), func(ctx context.Context) {
peer.Servers.Run(ctx, group)
peer.Services.Run(ctx, group)
pprof.Do(ctx, pprof.Labels("name", "subsystem-wait"), func(ctx context.Context) {
err = group.Wait()
})
})
return err
}
// Close closes all the resources.
func (peer *RangedLoop) Close() error {
return errs.Combine(
peer.Servers.Close(),
peer.Services.Close(),
)
}