storj/satellite/satellitedb/projectaccounting.go

628 lines
20 KiB
Go
Raw Normal View History

// Copyright (C) 2019 Storj Labs, Inc.
// See LICENSE for copying information.
package satellitedb
import (
"context"
"database/sql"
"fmt"
"time"
"github.com/lib/pq"
"github.com/skyrings/skyring-common/tools/uuid"
"github.com/zeebo/errs"
"storj.io/common/memory"
"storj.io/common/pb"
"storj.io/storj/private/dbutil"
"storj.io/storj/satellite/accounting"
"storj.io/storj/satellite/satellitedb/dbx"
)
// ensure that ProjectAccounting implements accounting.ProjectAccounting.
var _ accounting.ProjectAccounting = (*ProjectAccounting)(nil)
// ProjectAccounting implements the accounting/db ProjectAccounting interface
type ProjectAccounting struct {
satellite/satellitedb: unexport satellitedb.DB Backstory: I needed a better way to pass around information about the underlying driver and implementation to all the various db-using things in satellitedb (at least until some new "cockroach driver" support makes it to DBX). After hitting a few dead ends, I decided I wanted to have a type that could act like a *dbx.DB but which would also carry information about the implementation, etc. Then I could pass around that type to all the things in satellitedb that previously wanted *dbx.DB. But then I realized that *satellitedb.DB was, essentially, exactly that already. One thing that might have kept *satellitedb.DB from being directly usable was that embedding a *dbx.DB inside it would make a lot of dbx methods publicly available on a *satellitedb.DB instance that previously were nicely encapsulated and hidden. But after a quick look, I realized that _nothing_ outside of satellite/satellitedb even needs to use satellitedb.DB at all. It didn't even need to be exported, except for some trivially-replaceable code in migrate_postgres_test.go. And once I made it unexported, any concerns about exposing new methods on it were entirely moot. So I have here changed the exported *satellitedb.DB type into the unexported *satellitedb.satelliteDB type, and I have changed all the places here that wanted raw dbx.DB handles to use this new type instead. Now they can just take a gander at the implementation member on it and know all they need to know about the underlying database. This will make it possible for some other pending code here to differentiate between postgres and cockroach backends. Change-Id: I27af99f8ae23b50782333da5277b553b34634edc
2019-12-14 02:29:54 +00:00
db *satelliteDB
}
// SaveTallies saves the latest bucket info
func (db *ProjectAccounting) SaveTallies(ctx context.Context, intervalStart time.Time, bucketTallies map[string]*accounting.BucketTally) (err error) {
defer mon.Task()(&ctx)(&err)
if len(bucketTallies) == 0 {
return nil
}
var bucketNames, projectIDs [][]byte
var inlineBytes, remoteBytes, metadataSizes []uint64
var remoteSegments, inlineSegments, objectCounts []uint
for _, info := range bucketTallies {
bucketNames = append(bucketNames, info.BucketName)
projectIDs = append(projectIDs, info.ProjectID[:])
inlineBytes = append(inlineBytes, uint64(info.InlineBytes))
remoteBytes = append(remoteBytes, uint64(info.RemoteBytes))
remoteSegments = append(remoteSegments, uint(info.RemoteSegments))
inlineSegments = append(inlineSegments, uint(info.InlineSegments))
objectCounts = append(objectCounts, uint(info.ObjectCount))
metadataSizes = append(metadataSizes, uint64(info.MetadataSize))
}
_, err = db.db.DB.ExecContext(ctx, db.db.Rebind(`
INSERT INTO bucket_storage_tallies (
interval_start,
bucket_name, project_id,
inline, remote,
remote_segments_count, inline_segments_count,
object_count, metadata_size)
SELECT
$1,
unnest($2::bytea[]), unnest($3::bytea[]),
unnest($4::int8[]), unnest($5::int8[]),
unnest($6::int[]), unnest($7::int[]),
unnest($8::int[]), unnest($9::int8[])`),
intervalStart,
pq.ByteaArray(bucketNames), pq.ByteaArray(projectIDs),
pq.Array(inlineBytes), pq.Array(remoteBytes),
pq.Array(remoteSegments), pq.Array(inlineSegments),
pq.Array(objectCounts), pq.Array(metadataSizes))
return Error.Wrap(err)
}
// GetTallies saves the latest bucket info
func (db *ProjectAccounting) GetTallies(ctx context.Context) (tallies []accounting.BucketTally, err error) {
defer mon.Task()(&ctx)(&err)
dbxTallies, err := db.db.All_BucketStorageTally(ctx)
if err != nil {
return nil, Error.Wrap(err)
}
for _, dbxTally := range dbxTallies {
projectID, err := dbutil.BytesToUUID(dbxTally.ProjectId)
if err != nil {
return nil, Error.Wrap(err)
}
tallies = append(tallies, accounting.BucketTally{
BucketName: dbxTally.BucketName,
ProjectID: projectID,
ObjectCount: int64(dbxTally.ObjectCount),
InlineSegments: int64(dbxTally.InlineSegmentsCount),
RemoteSegments: int64(dbxTally.RemoteSegmentsCount),
InlineBytes: int64(dbxTally.Inline),
RemoteBytes: int64(dbxTally.Remote),
MetadataSize: int64(dbxTally.MetadataSize),
})
}
return tallies, nil
}
// CreateStorageTally creates a record in the bucket_storage_tallies accounting table
func (db *ProjectAccounting) CreateStorageTally(ctx context.Context, tally accounting.BucketStorageTally) (err error) {
defer mon.Task()(&ctx)(&err)
return Error.Wrap(db.db.CreateNoReturn_BucketStorageTally(
ctx,
dbx.BucketStorageTally_BucketName([]byte(tally.BucketName)),
dbx.BucketStorageTally_ProjectId(tally.ProjectID[:]),
dbx.BucketStorageTally_IntervalStart(tally.IntervalStart),
dbx.BucketStorageTally_Inline(uint64(tally.InlineBytes)),
dbx.BucketStorageTally_Remote(uint64(tally.RemoteBytes)),
dbx.BucketStorageTally_RemoteSegmentsCount(uint(tally.RemoteSegmentCount)),
dbx.BucketStorageTally_InlineSegmentsCount(uint(tally.InlineSegmentCount)),
dbx.BucketStorageTally_ObjectCount(uint(tally.ObjectCount)),
dbx.BucketStorageTally_MetadataSize(uint64(tally.MetadataSize)),
))
}
// GetAllocatedBandwidthTotal returns the sum of GET bandwidth usage allocated for a projectID for a time frame
func (db *ProjectAccounting) GetAllocatedBandwidthTotal(ctx context.Context, projectID uuid.UUID, from time.Time) (_ int64, err error) {
defer mon.Task()(&ctx)(&err)
var sum *int64
query := `SELECT SUM(allocated) FROM bucket_bandwidth_rollups WHERE project_id = ? AND action = ? AND interval_start > ?;`
err = db.db.QueryRow(ctx, db.db.Rebind(query), projectID[:], pb.PieceAction_GET, from).Scan(&sum)
if err == sql.ErrNoRows || sum == nil {
return 0, nil
}
return *sum, err
}
// GetStorageTotals returns the current inline and remote storage usage for a projectID
func (db *ProjectAccounting) GetStorageTotals(ctx context.Context, projectID uuid.UUID) (inline int64, remote int64, err error) {
defer mon.Task()(&ctx)(&err)
var inlineSum, remoteSum sql.NullInt64
var intervalStart time.Time
// Sum all the inline and remote values for a project that all share the same interval_start.
// All records for a project that have the same interval start are part of the same tally run.
// This should represent the most recent calculation of a project's total at rest storage.
query := `SELECT interval_start, SUM(inline), SUM(remote)
FROM bucket_storage_tallies
WHERE project_id = ?
GROUP BY interval_start
ORDER BY interval_start DESC LIMIT 1;`
err = db.db.QueryRow(ctx, db.db.Rebind(query), projectID[:]).Scan(&intervalStart, &inlineSum, &remoteSum)
if err != nil || !inlineSum.Valid || !remoteSum.Valid {
return 0, 0, nil
}
return inlineSum.Int64, remoteSum.Int64, err
}
// UpdateProjectUsageLimit updates project usage limit.
func (db *ProjectAccounting) UpdateProjectUsageLimit(ctx context.Context, projectID uuid.UUID, limit memory.Size) (err error) {
defer mon.Task()(&ctx)(&err)
_, err = db.db.Update_Project_By_Id(ctx,
dbx.Project_Id(projectID[:]),
dbx.Project_Update_Fields{
UsageLimit: dbx.Project_UsageLimit(limit.Int64()),
},
)
return err
}
// GetProjectStorageLimit returns project storage usage limit.
func (db *ProjectAccounting) GetProjectStorageLimit(ctx context.Context, projectID uuid.UUID) (_ memory.Size, err error) {
defer mon.Task()(&ctx)(&err)
return db.getProjectUsageLimit(ctx, projectID)
}
// GetProjectBandwidthLimit returns project bandwidth usage limit.
func (db *ProjectAccounting) GetProjectBandwidthLimit(ctx context.Context, projectID uuid.UUID) (_ memory.Size, err error) {
defer mon.Task()(&ctx)(&err)
return db.getProjectUsageLimit(ctx, projectID)
}
// getProjectUsageLimit returns project usage limit.
func (db *ProjectAccounting) getProjectUsageLimit(ctx context.Context, projectID uuid.UUID) (_ memory.Size, err error) {
defer mon.Task()(&ctx)(&err)
row, err := db.db.Get_Project_UsageLimit_By_Id(ctx,
dbx.Project_Id(projectID[:]),
)
if err != nil {
return 0, err
}
return memory.Size(row.UsageLimit), nil
}
// GetProjectTotal retrieves project usage for a given period.
func (db *ProjectAccounting) GetProjectTotal(ctx context.Context, projectID uuid.UUID, since, before time.Time) (usage *accounting.ProjectUsage, err error) {
defer mon.Task()(&ctx)(&err)
since = timeTruncateDown(since)
bucketNames, err := db.getBuckets(ctx, projectID, since, before)
if err != nil {
return nil, err
}
storageQuery := db.db.Rebind(`
SELECT
bucket_storage_tallies.interval_start,
bucket_storage_tallies.inline,
bucket_storage_tallies.remote,
bucket_storage_tallies.object_count
FROM
bucket_storage_tallies
WHERE
bucket_storage_tallies.project_id = ? AND
bucket_storage_tallies.bucket_name = ? AND
bucket_storage_tallies.interval_start >= ? AND
bucket_storage_tallies.interval_start <= ?
ORDER BY bucket_storage_tallies.interval_start DESC
`)
bucketsTallies := make(map[string][]*accounting.BucketStorageTally)
for _, bucket := range bucketNames {
storageTallies := make([]*accounting.BucketStorageTally, 0)
storageTalliesRows, err := db.db.QueryContext(ctx, storageQuery, projectID[:], []byte(bucket), since, before)
if err != nil {
return nil, err
}
// generating tallies for each bucket name.
for storageTalliesRows.Next() {
tally := accounting.BucketStorageTally{}
err = storageTalliesRows.Scan(&tally.IntervalStart, &tally.InlineBytes, &tally.RemoteBytes, &tally.ObjectCount)
if err != nil {
return nil, errs.Combine(err, storageTalliesRows.Close())
}
tally.BucketName = bucket
2019-11-28 21:42:04 +00:00
storageTallies = append(storageTallies, &tally)
}
err = errs.Combine(storageTalliesRows.Err(), storageTalliesRows.Close())
if err != nil {
return nil, err
}
bucketsTallies[bucket] = storageTallies
}
totalEgress, err := db.getTotalEgress(ctx, projectID, since, before)
if err != nil {
return nil, err
}
usage = new(accounting.ProjectUsage)
usage.Egress = memory.Size(totalEgress).Int64()
// sum up storage and objects
for _, tallies := range bucketsTallies {
for i := len(tallies) - 1; i > 0; i-- {
current := (tallies)[i]
hours := (tallies)[i-1].IntervalStart.Sub(current.IntervalStart).Hours()
usage.Storage += memory.Size(current.InlineBytes).Float64() * hours
usage.Storage += memory.Size(current.RemoteBytes).Float64() * hours
usage.ObjectCount += float64(current.ObjectCount) * hours
}
}
usage.Since = since
usage.Before = before
return usage, nil
}
// getTotalEgress returns total egress (settled + inline) of each bucket_bandwidth_rollup
// in selected time period, project id.
// only process PieceAction_GET.
func (db *ProjectAccounting) getTotalEgress(ctx context.Context, projectID uuid.UUID, since, before time.Time) (totalEgress int64, err error) {
totalEgressQuery := db.db.Rebind(`
SELECT
COALESCE(SUM(settled) + SUM(inline), 0)
FROM
bucket_bandwidth_rollups
WHERE
project_id = ? AND
interval_start >= ? AND
interval_start <= ? AND
action = ?;
`)
totalEgressRow := db.db.QueryRowContext(ctx, totalEgressQuery, projectID[:], since, before, pb.PieceAction_GET)
err = totalEgressRow.Scan(&totalEgress)
return totalEgress, err
}
// GetBucketUsageRollups retrieves summed usage rollups for every bucket of particular project for a given period
func (db *ProjectAccounting) GetBucketUsageRollups(ctx context.Context, projectID uuid.UUID, since, before time.Time) (_ []accounting.BucketUsageRollup, err error) {
defer mon.Task()(&ctx)(&err)
since = timeTruncateDown(since.UTC())
before = before.UTC()
buckets, err := db.getBuckets(ctx, projectID, since, before)
if err != nil {
return nil, err
}
roullupsQuery := db.db.Rebind(`SELECT SUM(settled), SUM(inline), action
FROM bucket_bandwidth_rollups
WHERE project_id = ? AND bucket_name = ? AND interval_start >= ? AND interval_start <= ?
GROUP BY action`)
// TODO: should be optimized
storageQuery := db.db.All_BucketStorageTally_By_ProjectId_And_BucketName_And_IntervalStart_GreaterOrEqual_And_IntervalStart_LessOrEqual_OrderBy_Desc_IntervalStart
var bucketUsageRollups []accounting.BucketUsageRollup
for _, bucket := range buckets {
bucketRollup := accounting.BucketUsageRollup{
ProjectID: projectID,
BucketName: []byte(bucket),
Since: since,
Before: before,
}
// get bucket_bandwidth_rollups
rollupsRows, err := db.db.QueryContext(ctx, roullupsQuery, projectID[:], []byte(bucket), since, before)
if err != nil {
return nil, err
}
defer func() { err = errs.Combine(err, rollupsRows.Close()) }()
// fill egress
for rollupsRows.Next() {
var action pb.PieceAction
var settled, inline int64
err = rollupsRows.Scan(&settled, &inline, &action)
if err != nil {
return nil, err
}
switch action {
case pb.PieceAction_GET:
bucketRollup.GetEgress += memory.Size(settled + inline).GB()
case pb.PieceAction_GET_AUDIT:
bucketRollup.AuditEgress += memory.Size(settled + inline).GB()
case pb.PieceAction_GET_REPAIR:
bucketRollup.RepairEgress += memory.Size(settled + inline).GB()
default:
continue
}
}
if err := rollupsRows.Err(); err != nil {
return nil, err
}
bucketStorageTallies, err := storageQuery(ctx,
dbx.BucketStorageTally_ProjectId(projectID[:]),
dbx.BucketStorageTally_BucketName([]byte(bucket)),
dbx.BucketStorageTally_IntervalStart(since),
dbx.BucketStorageTally_IntervalStart(before))
if err != nil {
return nil, err
}
// fill metadata, objects and stored data
// hours calculated from previous tallies,
// so we skip the most recent one
for i := len(bucketStorageTallies) - 1; i > 0; i-- {
current := bucketStorageTallies[i]
hours := bucketStorageTallies[i-1].IntervalStart.Sub(current.IntervalStart).Hours()
bucketRollup.RemoteStoredData += memory.Size(current.Remote).GB() * hours
bucketRollup.InlineStoredData += memory.Size(current.Inline).GB() * hours
bucketRollup.MetadataSize += memory.Size(current.MetadataSize).GB() * hours
bucketRollup.RemoteSegments += float64(current.RemoteSegmentsCount) * hours
bucketRollup.InlineSegments += float64(current.InlineSegmentsCount) * hours
bucketRollup.ObjectCount += float64(current.ObjectCount) * hours
}
bucketUsageRollups = append(bucketUsageRollups, bucketRollup)
}
return bucketUsageRollups, nil
}
// prefixIncrement returns the lexicographically lowest byte string which is
// greater than origPrefix and does not have origPrefix as a prefix. If no such
// byte string exists (origPrefix is empty, or origPrefix contains only 0xff
// bytes), returns false for ok.
//
// examples: prefixIncrement([]byte("abc")) -> ([]byte("abd", true)
// prefixIncrement([]byte("ab\xff\xff")) -> ([]byte("ac", true)
// prefixIncrement([]byte("")) -> (nil, false)
// prefixIncrement([]byte("\x00")) -> ([]byte("\x01", true)
// prefixIncrement([]byte("\xff\xff\xff")) -> (nil, false)
//
func prefixIncrement(origPrefix []byte) (incremented []byte, ok bool) {
incremented = make([]byte, len(origPrefix))
copy(incremented, origPrefix)
i := len(incremented) - 1
for i >= 0 {
if incremented[i] != 0xff {
incremented[i]++
return incremented[:i+1], true
}
i--
}
// there is no byte string which is greater than origPrefix and which does
// not have origPrefix as a prefix.
return nil, false
}
// prefixMatch creates a SQL expression which
// will evaluate to true if and only if the value of expr starts with the value
// of prefix.
//
// Returns also a slice of arguments that should be passed to the corresponding
// db.Query* or db.Exec* to fill in parameters in the returned SQL expression.
//
// The returned SQL expression needs to be passed through Rebind(), as it uses
// `?` markers instead of `$N`, because we don't know what N we would need to
// use.
func (db *ProjectAccounting) prefixMatch(expr string, prefix []byte) (string, []byte, error) {
incrementedPrefix, ok := prefixIncrement(prefix)
switch db.db.implementation {
case dbutil.Postgres:
if !ok {
return fmt.Sprintf(`(%s >= ?)`, expr), nil, nil
}
return fmt.Sprintf(`(%s >= ? AND %s < ?)`, expr, expr), incrementedPrefix, nil
case dbutil.Cockroach:
if !ok {
return fmt.Sprintf(`(%s >= ?:::BYTEA)`, expr), nil, nil
}
return fmt.Sprintf(`(%s >= ?:::BYTEA AND %s < ?:::BYTEA)`, expr, expr), incrementedPrefix, nil
default:
return "", nil, errs.New("invalid dbType: %v", db.db.driver)
}
}
// GetBucketTotals retrieves bucket usage totals for period of time
func (db *ProjectAccounting) GetBucketTotals(ctx context.Context, projectID uuid.UUID, cursor accounting.BucketUsageCursor, since, before time.Time) (_ *accounting.BucketUsagePage, err error) {
defer mon.Task()(&ctx)(&err)
since = timeTruncateDown(since)
bucketPrefix := []byte(cursor.Search)
if cursor.Limit > 50 {
cursor.Limit = 50
}
if cursor.Page == 0 {
return nil, errs.New("page can not be 0")
}
page := &accounting.BucketUsagePage{
Search: cursor.Search,
Limit: cursor.Limit,
Offset: uint64((cursor.Page - 1) * cursor.Limit),
}
bucketNameRange, incrPrefix, err := db.prefixMatch("bucket_name", bucketPrefix)
if err != nil {
return nil, err
}
countQuery := db.db.Rebind(`SELECT COUNT(DISTINCT bucket_name)
FROM bucket_bandwidth_rollups
WHERE project_id = ? AND interval_start >= ? AND interval_start <= ?
AND ` + bucketNameRange)
args := []interface{}{
projectID[:],
since,
before,
bucketPrefix,
}
if incrPrefix != nil {
args = append(args, incrPrefix)
}
countRow := db.db.QueryRowContext(ctx, countQuery, args...)
err = countRow.Scan(&page.TotalCount)
if err != nil {
return nil, err
}
if page.TotalCount == 0 {
return page, nil
}
if page.Offset > page.TotalCount-1 {
return nil, errs.New("page is out of range")
}
var buckets []string
bucketsQuery := db.db.Rebind(`SELECT DISTINCT bucket_name
FROM bucket_bandwidth_rollups
WHERE project_id = ? AND interval_start >= ? AND interval_start <= ?
AND ` + bucketNameRange + ` ORDER BY bucket_name ASC
LIMIT ? OFFSET ?`)
args = []interface{}{
projectID[:],
since,
before,
bucketPrefix,
}
if incrPrefix != nil {
args = append(args, incrPrefix)
}
args = append(args, page.Limit, page.Offset)
bucketRows, err := db.db.QueryContext(ctx, bucketsQuery, args...)
if err != nil {
return nil, err
}
defer func() { err = errs.Combine(err, bucketRows.Close()) }()
for bucketRows.Next() {
var bucket string
err = bucketRows.Scan(&bucket)
if err != nil {
return nil, err
}
buckets = append(buckets, bucket)
}
if err := bucketRows.Err(); err != nil {
return nil, err
}
rollupsQuery := db.db.Rebind(`SELECT COALESCE(SUM(settled) + SUM(inline), 0)
FROM bucket_bandwidth_rollups
WHERE project_id = ? AND bucket_name = ? AND interval_start >= ? AND interval_start <= ? AND action = ?`)
storageQuery := db.db.Rebind(`SELECT inline, remote, object_count
FROM bucket_storage_tallies
WHERE project_id = ? AND bucket_name = ? AND interval_start >= ? AND interval_start <= ?
ORDER BY interval_start DESC
LIMIT 1`)
var bucketUsages []accounting.BucketUsage
for _, bucket := range buckets {
bucketUsage := accounting.BucketUsage{
ProjectID: projectID,
BucketName: bucket,
Since: since,
Before: before,
}
// get bucket_bandwidth_rollups
rollupRow := db.db.QueryRowContext(ctx, rollupsQuery, projectID[:], []byte(bucket), since, before, pb.PieceAction_GET)
var egress int64
err = rollupRow.Scan(&egress)
if err != nil {
if err != sql.ErrNoRows {
return nil, err
}
}
bucketUsage.Egress = memory.Size(egress).GB()
storageRow := db.db.QueryRowContext(ctx, storageQuery, projectID[:], []byte(bucket), since, before)
var inline, remote, objectCount int64
err = storageRow.Scan(&inline, &remote, &objectCount)
if err != nil {
if err != sql.ErrNoRows {
return nil, err
}
}
// fill storage and object count
bucketUsage.Storage = memory.Size(inline + remote).GB()
bucketUsage.ObjectCount = objectCount
bucketUsages = append(bucketUsages, bucketUsage)
}
page.PageCount = uint(page.TotalCount / uint64(cursor.Limit))
if page.TotalCount%uint64(cursor.Limit) != 0 {
page.PageCount++
}
page.BucketUsages = bucketUsages
page.CurrentPage = cursor.Page
return page, nil
}
// getBuckets list all bucket of certain project for given period
func (db *ProjectAccounting) getBuckets(ctx context.Context, projectID uuid.UUID, since, before time.Time) (_ []string, err error) {
defer mon.Task()(&ctx)(&err)
bucketsQuery := db.db.Rebind(`SELECT DISTINCT bucket_name
FROM bucket_bandwidth_rollups
WHERE project_id = ? AND interval_start >= ? AND interval_start <= ?`)
bucketRows, err := db.db.QueryContext(ctx, bucketsQuery, projectID[:], since, before)
if err != nil {
return nil, err
}
defer func() { err = errs.Combine(err, bucketRows.Close()) }()
var buckets []string
for bucketRows.Next() {
var bucket string
err = bucketRows.Scan(&bucket)
if err != nil {
return nil, err
}
buckets = append(buckets, bucket)
}
return buckets, bucketRows.Err()
}
// timeTruncateDown truncates down to the hour before to be in sync with orders endpoint
func timeTruncateDown(t time.Time) time.Time {
return time.Date(t.Year(), t.Month(), t.Day(), t.Hour(), 0, 0, 0, t.Location())
}