storj/pkg/overlay/cache.go

234 lines
7.5 KiB
Go
Raw Normal View History

2019-01-24 20:15:10 +00:00
// Copyright (C) 2019 Storj Labs, Inc.
2018-04-18 17:55:28 +01:00
// See LICENSE for copying information.
package overlay
2018-04-18 16:34:15 +01:00
import (
"context"
"errors"
2018-04-18 16:34:15 +01:00
"github.com/zeebo/errs"
"go.uber.org/zap"
2018-11-16 16:31:14 +00:00
"storj.io/storj/pkg/pb"
"storj.io/storj/pkg/statdb"
2018-11-30 13:40:13 +00:00
"storj.io/storj/pkg/storj"
"storj.io/storj/storage"
2018-04-18 16:34:15 +01:00
)
const (
// OverlayBucket is the string representing the bucket used for a bolt-backed overlay dht cache
OverlayBucket = "overlay"
)
2018-12-17 18:47:26 +00:00
// ErrEmptyNode is returned when the nodeID is empty
var ErrEmptyNode = errs.New("empty node ID")
// ErrNodeNotFound is returned if a node does not exist in database
var ErrNodeNotFound = errs.New("Node not found")
2018-12-17 18:47:26 +00:00
// ErrBucketNotFound is returned if a bucket is unable to be found in the routing table
var ErrBucketNotFound = errs.New("Bucket not found")
// ErrNotEnoughNodes is when selecting nodes failed with the given parameters
var ErrNotEnoughNodes = errs.Class("not enough nodes")
// OverlayError creates class of errors for stack traces
var OverlayError = errs.Class("Overlay Error")
Cache (#67) * add reference to dht to overlay client struct * wip * wip * Implement FindNode * get nodes * WIP * Merge in Dennis kademlia code, get it working with our code * ping and moar * WIP trying to get cache working with kademlia * WIP more wiring up * WIP * Update service cli commands * WIP * added GetNodes * added nodes to Kbucket * default transport changed to TCP * GetBuckets interface changed * filling in more routing * timestamp methods * removed store * Added initial network overlay explorer page * Updating and building with dockerfile * Working on adding bootstrap node code * WIP merging in dennis' code * WIP * connects cache to pkg/kademlia implementation * WIP redis cache * testing * Add bootstrap network function for CLI usage * cleanup * call bootstrap on init network * Add BootstrapNetwork function to interface * Merge in dennis kad code * WIP updates to redis/overlay client interface * WIP trying to get the DHT connected to the cache * go mod & test * deps * Bootstrap node now setting up correctly - Need to pass it through CLI commands better * WIP adding refresh and walk functions, added cli flags - added cli flags for custom bootstrap port and ip * PR comments addressed * adding FindStorageNodes to overlay cache * fix GetBucket * using SplitHostPort * Use JoinHostPort * updates to findstoragenodes response and request * WIP merge in progress, having issues with a panic * wip * adjustments * update port for dht bootstrap test * Docker * wip * dockerfile * fixes * makefile changes * Update port in NewKademlia call * Update local kademlia DHT config * kubernetes yaml * cleanup * making tests pass * k8s yaml * lint issues * Edit cli flags to allow for configurable bootstrap IP and Port args * cleanup * cache walking the network now * Rough prototype of Walk function laid out * Move walk function into bootstrap function * Update dht.go * changes to yaml * goimports
2018-06-05 22:06:37 +01:00
// DB implements the database for overlay.Cache
type DB interface {
// SelectStorageNodes looks up nodes based on criteria
SelectStorageNodes(ctx context.Context, count int, criteria *NodeCriteria) ([]*pb.Node, error)
// SelectNewStorageNodes looks up nodes based on new node criteria
SelectNewStorageNodes(ctx context.Context, count int, criteria *NewNodeCriteria) ([]*pb.Node, error)
// Get looks up the node by nodeID
Get(ctx context.Context, nodeID storj.NodeID) (*pb.Node, error)
// GetAll looks up nodes based on the ids from the overlay cache
GetAll(ctx context.Context, nodeIDs storj.NodeIDList) ([]*pb.Node, error)
// List lists nodes starting from cursor
List(ctx context.Context, cursor storj.NodeID, limit int) ([]*pb.Node, error)
// Paginate will page through the database nodes
Paginate(ctx context.Context, offset int64, limit int) ([]*pb.Node, bool, error)
// Update updates node information
Update(ctx context.Context, value *pb.Node) error
// Delete deletes node based on id
Delete(ctx context.Context, id storj.NodeID) error
// GetWalletAddress gets the node's wallet address
2019-01-17 18:34:13 +00:00
GetWalletAddress(ctx context.Context, id storj.NodeID) (string, error)
}
// Cache is used to store overlay data in Redis
type Cache struct {
db DB
statDB statdb.DB
2018-04-18 16:34:15 +01:00
}
// NewCache returns a new Cache
func NewCache(db DB, sdb statdb.DB) *Cache {
return &Cache{db: db, statDB: sdb}
}
Satellite Peer (#1034) * add satellite peer * Add overlay * reorganize kademlia * add RunRefresh * add refresh to storagenode.Peer * add discovery * add agreements and metainfo * rename * add datarepair checker * add repair * add todo notes for audit * add testing interface * add into testplanet * fixes * fix compilation errors * fix compilation errors * make testplanet run * remove audit refrences * ensure that audit tests run * dev * checker tests compilable * fix discovery * fix compilation * fix * fix * dev * fix * disable auth * fixes * revert go.mod/sum * fix linter errors * fix * fix copyright * Add address param for SN dashboard (#1076) * Rename storj-sdk to storj-sim (#1078) * Storagenode logs and config improvements (#1075) * Add more info to SN logs * remove config-dir from user config * add output where config was stored * add message for successful connection * fix linter * remove storage.path from user config * resolve config path * move success message to info * log improvements * Remove captplanet (#1070) * pkg/server: include production cert (#1082) Change-Id: Ie8e6fe78550be83c3bd797db7a1e58d37c684792 * Generate Payments Report (#1079) * memory.Size: autoformat sizes based on value entropy (#1081) * Jj/bytes (#1085) * run tally and rollup * sets dev default tally and rollup intervals * nonessential storj-sim edits (#1086) * Closing context doesn't stop storage node (#1084) * Print when cancelled * Close properly * Don't log nil * Don't print error when closing dashboard * Fix panic in inspector if ping fails (#1088) * Consolidate identity management to identity cli commands (#1083) * Consolidate identity management: Move identity cretaion/signing out of storagenode setup command. * fixes * linters * Consolidate identity management: Move identity cretaion/signing out of storagenode setup command. * fixes * sava backups before saving signed certs * add "-prebuilt-test-cmds" test flag * linters * prepare cli tests for travis * linter fixes * more fixes * linter gods * sp/sdk/sim * remove ca.difficulty * remove unused difficulty * return setup to its rightful place * wip travis * Revert "wip travis" This reverts commit 56834849dcf066d3cc0a4f139033fc3f6d7188ca. * typo in travis.yaml * remove tests * remove more * make it only create one identity at a time for consistency * add config-dir for consitency * add identity creation to storj-sim * add flags * simplify * fix nolint and compile * prevent overwrite and pass difficulty, concurrency, and parent creds * goimports
2019-01-18 13:54:08 +00:00
// Close closes resources
func (cache *Cache) Close() error { return nil }
// Inspect lists limited number of items in the cache
func (cache *Cache) Inspect(ctx context.Context) (storage.Keys, error) {
// TODO: implement inspection tools
return nil, errors.New("not implemented")
}
// List returns a list of nodes from the cache DB
func (cache *Cache) List(ctx context.Context, cursor storj.NodeID, limit int) ([]*pb.Node, error) {
return cache.db.List(ctx, cursor, limit)
}
// Paginate returns a list of `limit` nodes starting from `start` offset.
func (cache *Cache) Paginate(ctx context.Context, offset int64, limit int) ([]*pb.Node, bool, error) {
return cache.db.Paginate(ctx, offset, limit)
}
// Get looks up the provided nodeID from the overlay cache
func (cache *Cache) Get(ctx context.Context, nodeID storj.NodeID) (*pb.Node, error) {
2018-12-17 18:47:26 +00:00
if nodeID.IsZero() {
return nil, ErrEmptyNode
}
return cache.db.Get(ctx, nodeID)
2018-04-18 16:34:15 +01:00
}
// FindStorageNodes searches the overlay network for nodes that meet the provided criteria
func (cache *Cache) FindStorageNodes(ctx context.Context, req *pb.FindStorageNodesRequest, preferences *NodeSelectionConfig) ([]*pb.Node, error) {
// TODO: use a nicer struct for input
minimumRequiredNodes := int(req.GetMinNodes())
freeBandwidth := req.GetOpts().GetRestrictions().FreeBandwidth
freeDisk := req.GetOpts().GetRestrictions().FreeDisk
excludedNodes := req.GetOpts().ExcludedNodes
requestedCount := int(req.GetOpts().GetAmount())
// TODO: verify logic
// TODO: add sanity limits to requested node count
// TODO: add sanity limits to excluded nodes
reputableNodeCount := minimumRequiredNodes
if reputableNodeCount <= 0 {
reputableNodeCount = requestedCount
}
auditCount := preferences.AuditCount
if auditCount < preferences.NewNodeAuditThreshold {
auditCount = preferences.NewNodeAuditThreshold
}
reputableNodes, err := cache.db.SelectStorageNodes(ctx, reputableNodeCount, &NodeCriteria{
FreeBandwidth: freeBandwidth,
FreeDisk: freeDisk,
AuditCount: auditCount,
AuditSuccessRatio: preferences.AuditSuccessRatio,
UptimeCount: preferences.UptimeCount,
UptimeSuccessRatio: preferences.UptimeRatio,
Excluded: excludedNodes,
})
if err != nil {
return nil, err
}
newNodeCount := int64(float64(reputableNodeCount) * preferences.NewNodePercentage)
newNodes, err := cache.db.SelectNewStorageNodes(ctx, int(newNodeCount), &NewNodeCriteria{
FreeBandwidth: freeBandwidth,
FreeDisk: freeDisk,
AuditThreshold: preferences.NewNodeAuditThreshold,
Excluded: excludedNodes,
})
if err != nil {
return nil, err
}
nodes := []*pb.Node{}
nodes = append(nodes, newNodes...)
nodes = append(nodes, reputableNodes...)
if len(reputableNodes) < reputableNodeCount {
return nodes, ErrNotEnoughNodes.New("requested %d found %d", reputableNodeCount, len(reputableNodes))
}
return nodes, nil
}
// GetAll looks up the provided ids from the overlay cache
func (cache *Cache) GetAll(ctx context.Context, ids storj.NodeIDList) ([]*pb.Node, error) {
if len(ids) == 0 {
return nil, OverlayError.New("no ids provided")
}
return cache.db.GetAll(ctx, ids)
}
// Put adds a nodeID to the redis cache with a binary representation of proto defined Node
func (cache *Cache) Put(ctx context.Context, nodeID storj.NodeID, value pb.Node) error {
// If we get a Node without an ID (i.e. bootstrap node)
// we don't want to add to the routing tbale
2018-12-17 18:47:26 +00:00
if nodeID.IsZero() {
return nil
}
if nodeID != value.Id {
return errors.New("invalid request")
}
// get existing node rep, or create a new statdb node with 0 rep
stats, err := cache.statDB.CreateEntryIfNotExists(ctx, nodeID)
if err != nil {
return err
}
value.Reputation = &pb.NodeStats{
AuditSuccessRatio: stats.AuditSuccessRatio,
AuditSuccessCount: stats.AuditSuccessCount,
AuditCount: stats.AuditCount,
UptimeRatio: stats.UptimeRatio,
UptimeSuccessCount: stats.UptimeSuccessCount,
UptimeCount: stats.UptimeCount,
}
return cache.db.Update(ctx, &value)
2018-04-18 16:34:15 +01:00
}
// Delete will remove the node from the cache. Used when a node hard disconnects or fails
// to pass a PING multiple times.
func (cache *Cache) Delete(ctx context.Context, id storj.NodeID) error {
if id.IsZero() {
return ErrEmptyNode
}
return cache.db.Delete(ctx, id)
}
// ConnFailure implements the Transport Observer `ConnFailure` function
func (cache *Cache) ConnFailure(ctx context.Context, node *pb.Node, failureError error) {
// TODO: Kademlia paper specifies 5 unsuccessful PINGs before removing the node
// from our routing table, but this is the cache so maybe we want to treat
// it differently.
_, err := cache.statDB.UpdateUptime(ctx, node.Id, false)
if err != nil {
zap.L().Debug("error updating uptime for node in statDB", zap.Error(err))
}
}
// ConnSuccess implements the Transport Observer `ConnSuccess` function
func (cache *Cache) ConnSuccess(ctx context.Context, node *pb.Node) {
err := cache.Put(ctx, node.Id, *node)
if err != nil {
zap.L().Debug("error updating uptime for node in statDB", zap.Error(err))
}
_, err = cache.statDB.UpdateUptime(ctx, node.Id, true)
if err != nil {
zap.L().Debug("error updating statdDB with node connection info", zap.Error(err))
}
}