forked from TrueCloudLab/frostfs-node
[#1004] blobovnicza: Use TTL for blobovnicza tree cache
Signed-off-by: Anton Nikiforov <an.nikiforov@yadro.com>
This commit is contained in:
parent
112a7c690f
commit
411a8d0245
26 changed files with 203 additions and 65 deletions
|
@ -61,6 +61,8 @@ storage:
|
|||
depth: 1 # max depth of object tree storage in key-value DB
|
||||
width: 4 # max width of object tree storage in key-value DB
|
||||
opened_cache_capacity: 50 # maximum number of opened database files
|
||||
opened_cache_ttl: 5m # ttl for opened database file
|
||||
opened_cache_exp_interval: 15s # cache cleanup interval for expired blobovnicza's
|
||||
|
||||
gc:
|
||||
remover_batch_size: 200 # number of objects to be removed by the garbage collector
|
||||
|
|
|
@ -185,13 +185,15 @@ type subStorageCfg struct {
|
|||
noSync bool
|
||||
|
||||
// blobovnicza-specific
|
||||
size uint64
|
||||
width uint64
|
||||
leafWidth uint64
|
||||
openedCacheSize int
|
||||
initWorkerCount int
|
||||
initInAdvance bool
|
||||
rebuildDropTimeout time.Duration
|
||||
size uint64
|
||||
width uint64
|
||||
leafWidth uint64
|
||||
openedCacheSize int
|
||||
initWorkerCount int
|
||||
initInAdvance bool
|
||||
rebuildDropTimeout time.Duration
|
||||
openedCacheTTL time.Duration
|
||||
openedCacheExpInterval time.Duration
|
||||
}
|
||||
|
||||
// readConfig fills applicationConfiguration with raw configuration values
|
||||
|
@ -313,6 +315,8 @@ func (a *applicationConfiguration) setShardStorageConfig(newConfig *shardCfg, ol
|
|||
sCfg.width = sub.ShallowWidth()
|
||||
sCfg.leafWidth = sub.LeafWidth()
|
||||
sCfg.openedCacheSize = sub.OpenedCacheSize()
|
||||
sCfg.openedCacheTTL = sub.OpenedCacheTTL()
|
||||
sCfg.openedCacheExpInterval = sub.OpenedCacheExpInterval()
|
||||
sCfg.initWorkerCount = sub.InitWorkerCount()
|
||||
sCfg.initInAdvance = sub.InitInAdvance()
|
||||
sCfg.rebuildDropTimeout = sub.RebuildDropTimeout()
|
||||
|
@ -846,11 +850,11 @@ type shardOptsWithID struct {
|
|||
shOpts []shard.Option
|
||||
}
|
||||
|
||||
func (c *cfg) shardOpts() []shardOptsWithID {
|
||||
func (c *cfg) shardOpts(ctx context.Context) []shardOptsWithID {
|
||||
shards := make([]shardOptsWithID, 0, len(c.EngineCfg.shards))
|
||||
|
||||
for _, shCfg := range c.EngineCfg.shards {
|
||||
shards = append(shards, c.getShardOpts(shCfg))
|
||||
shards = append(shards, c.getShardOpts(ctx, shCfg))
|
||||
}
|
||||
|
||||
return shards
|
||||
|
@ -891,7 +895,7 @@ func (c *cfg) getPiloramaOpts(shCfg shardCfg) []pilorama.Option {
|
|||
return piloramaOpts
|
||||
}
|
||||
|
||||
func (c *cfg) getSubstorageOpts(shCfg shardCfg) []blobstor.SubStorage {
|
||||
func (c *cfg) getSubstorageOpts(ctx context.Context, shCfg shardCfg) []blobstor.SubStorage {
|
||||
var ss []blobstor.SubStorage
|
||||
for _, sRead := range shCfg.subStorages {
|
||||
switch sRead.typ {
|
||||
|
@ -904,6 +908,8 @@ func (c *cfg) getSubstorageOpts(shCfg shardCfg) []blobstor.SubStorage {
|
|||
blobovniczatree.WithBlobovniczaShallowWidth(sRead.width),
|
||||
blobovniczatree.WithBlobovniczaLeafWidth(sRead.leafWidth),
|
||||
blobovniczatree.WithOpenedCacheSize(sRead.openedCacheSize),
|
||||
blobovniczatree.WithOpenedCacheTTL(sRead.openedCacheTTL),
|
||||
blobovniczatree.WithOpenedCacheExpInterval(sRead.openedCacheExpInterval),
|
||||
blobovniczatree.WithInitWorkerCount(sRead.initWorkerCount),
|
||||
blobovniczatree.WithInitInAdvance(sRead.initInAdvance),
|
||||
blobovniczatree.WithWaitBeforeDropDB(sRead.rebuildDropTimeout),
|
||||
|
@ -919,7 +925,7 @@ func (c *cfg) getSubstorageOpts(shCfg shardCfg) []blobstor.SubStorage {
|
|||
)
|
||||
}
|
||||
ss = append(ss, blobstor.SubStorage{
|
||||
Storage: blobovniczatree.NewBlobovniczaTree(blobTreeOpts...),
|
||||
Storage: blobovniczatree.NewBlobovniczaTree(ctx, blobTreeOpts...),
|
||||
Policy: func(_ *objectSDK.Object, data []byte) bool {
|
||||
return uint64(len(data)) < shCfg.smallSizeObjectLimit
|
||||
},
|
||||
|
@ -954,10 +960,10 @@ func (c *cfg) getSubstorageOpts(shCfg shardCfg) []blobstor.SubStorage {
|
|||
return ss
|
||||
}
|
||||
|
||||
func (c *cfg) getShardOpts(shCfg shardCfg) shardOptsWithID {
|
||||
func (c *cfg) getShardOpts(ctx context.Context, shCfg shardCfg) shardOptsWithID {
|
||||
writeCacheOpts := c.getWriteCacheOpts(shCfg)
|
||||
piloramaOpts := c.getPiloramaOpts(shCfg)
|
||||
ss := c.getSubstorageOpts(shCfg)
|
||||
ss := c.getSubstorageOpts(ctx, shCfg)
|
||||
|
||||
blobstoreOpts := []blobstor.Option{
|
||||
blobstor.WithCompressObjects(shCfg.compress),
|
||||
|
@ -1049,7 +1055,7 @@ func initLocalStorage(ctx context.Context, c *cfg) {
|
|||
c.cfgObject.getSvc = new(getsvc.Service)
|
||||
|
||||
var shardsAttached int
|
||||
for _, optsWithMeta := range c.shardOpts() {
|
||||
for _, optsWithMeta := range c.shardOpts(ctx) {
|
||||
id, err := ls.AddShard(ctx, append(optsWithMeta.shOpts, shard.WithTombstoneSource(c.createTombstoneSource()))...)
|
||||
if err != nil {
|
||||
c.log.Error(logs.FrostFSNodeFailedToAttachShardToEngine, zap.Error(err))
|
||||
|
@ -1289,7 +1295,7 @@ func (c *cfg) reloadConfig(ctx context.Context) {
|
|||
// Storage Engine
|
||||
|
||||
var rcfg engine.ReConfiguration
|
||||
for _, optsWithID := range c.shardOpts() {
|
||||
for _, optsWithID := range c.shardOpts(ctx) {
|
||||
rcfg.AddShard(optsWithID.configID, append(optsWithID.shOpts, shard.WithTombstoneSource(c.createTombstoneSource())))
|
||||
}
|
||||
|
||||
|
|
|
@ -25,6 +25,12 @@ const (
|
|||
// OpenedCacheSizeDefault is a default cache size of opened Blobovnicza's.
|
||||
OpenedCacheSizeDefault = 16
|
||||
|
||||
// OpenedCacheTTLDefault is a default cache ttl of opened Blobovnicza's.
|
||||
OpenedCacheTTLDefault = 0 // means expiring is off
|
||||
|
||||
// OpenedCacheExpIntervalDefault is a default cache cleanup interval for expired Blobovnicza's.
|
||||
OpenedCacheExpIntervalDefault = 15 * time.Second
|
||||
|
||||
// InitWorkerCountDefault is a default workers count to initialize Blobovnicza's.
|
||||
InitWorkerCountDefault = 5
|
||||
|
||||
|
@ -106,6 +112,38 @@ func (x *Config) OpenedCacheSize() int {
|
|||
return OpenedCacheSizeDefault
|
||||
}
|
||||
|
||||
// OpenedCacheTTL returns the value of "opened_cache_ttl" config parameter.
|
||||
//
|
||||
// Returns OpenedCacheTTLDefault if the value is not a positive number.
|
||||
func (x *Config) OpenedCacheTTL() time.Duration {
|
||||
d := config.DurationSafe(
|
||||
(*config.Config)(x),
|
||||
"opened_cache_ttl",
|
||||
)
|
||||
|
||||
if d > 0 {
|
||||
return d
|
||||
}
|
||||
|
||||
return OpenedCacheTTLDefault
|
||||
}
|
||||
|
||||
// OpenedCacheExpInterval returns the value of "opened_cache_exp_interval" config parameter.
|
||||
//
|
||||
// Returns OpenedCacheExpIntervalDefault if the value is not a positive number.
|
||||
func (x *Config) OpenedCacheExpInterval() time.Duration {
|
||||
d := config.DurationSafe(
|
||||
(*config.Config)(x),
|
||||
"opened_cache_exp_interval",
|
||||
)
|
||||
|
||||
if d > 0 {
|
||||
return d
|
||||
}
|
||||
|
||||
return OpenedCacheExpIntervalDefault
|
||||
}
|
||||
|
||||
// BoltDB returns config instance for querying bolt db specific parameters.
|
||||
func (x *Config) BoltDB() *boltdbconfig.Config {
|
||||
return (*boltdbconfig.Config)(x)
|
||||
|
|
|
@ -128,6 +128,8 @@ FROSTFS_STORAGE_SHARD_0_BLOBSTOR_0_SIZE=4194304
|
|||
FROSTFS_STORAGE_SHARD_0_BLOBSTOR_0_DEPTH=1
|
||||
FROSTFS_STORAGE_SHARD_0_BLOBSTOR_0_WIDTH=4
|
||||
FROSTFS_STORAGE_SHARD_0_BLOBSTOR_0_OPENED_CACHE_CAPACITY=50
|
||||
FROSTFS_STORAGE_SHARD_0_BLOBSTOR_0_OPENED_CACHE_TTL=5m
|
||||
FROSTFS_STORAGE_SHARD_0_BLOBSTOR_0_OPENED_CACHE_EXP_INTERVAL=15s
|
||||
FROSTFS_STORAGE_SHARD_0_BLOBSTOR_0_LEAF_WIDTH=10
|
||||
FROSTFS_STORAGE_SHARD_0_BLOBSTOR_0_INIT_WORKER_COUNT=10
|
||||
FROSTFS_STORAGE_SHARD_0_BLOBSTOR_0_INIT_IN_ADVANCE=TRUE
|
||||
|
@ -178,6 +180,8 @@ FROSTFS_STORAGE_SHARD_1_BLOBSTOR_0_SIZE=4194304
|
|||
FROSTFS_STORAGE_SHARD_1_BLOBSTOR_0_DEPTH=1
|
||||
FROSTFS_STORAGE_SHARD_1_BLOBSTOR_0_WIDTH=4
|
||||
FROSTFS_STORAGE_SHARD_1_BLOBSTOR_0_OPENED_CACHE_CAPACITY=50
|
||||
FROSTFS_STORAGE_SHARD_1_BLOBSTOR_0_OPENED_CACHE_TTL=5m
|
||||
FROSTFS_STORAGE_SHARD_1_BLOBSTOR_0_OPENED_CACHE_EXP_INTERVAL=15s
|
||||
FROSTFS_STORAGE_SHARD_1_BLOBSTOR_0_LEAF_WIDTH=10
|
||||
### FSTree config
|
||||
FROSTFS_STORAGE_SHARD_1_BLOBSTOR_1_TYPE=fstree
|
||||
|
|
|
@ -176,6 +176,8 @@
|
|||
"depth": 1,
|
||||
"width": 4,
|
||||
"opened_cache_capacity": 50,
|
||||
"opened_cache_ttl": "5m",
|
||||
"opened_cache_exp_interval": "15s",
|
||||
"leaf_width": 10,
|
||||
"init_worker_count": 10,
|
||||
"init_in_advance": true,
|
||||
|
@ -229,6 +231,8 @@
|
|||
"depth": 1,
|
||||
"width": 4,
|
||||
"opened_cache_capacity": 50,
|
||||
"opened_cache_ttl": "5m",
|
||||
"opened_cache_exp_interval": "15s",
|
||||
"leaf_width": 10
|
||||
},
|
||||
{
|
||||
|
|
|
@ -151,6 +151,8 @@ storage:
|
|||
depth: 1 # max depth of object tree storage in key-value DB
|
||||
width: 4 # max width of object tree storage in key-value DB
|
||||
opened_cache_capacity: 50 # maximum number of opened database files
|
||||
opened_cache_ttl: 5m # ttl for opened database file
|
||||
opened_cache_exp_interval: 15s # cache cleanup interval for expired blobovnicza's
|
||||
leaf_width: 10 # max count of key-value DB on leafs of object tree storage
|
||||
- perm: 0644 # permissions for blobstor files(directories: +x for current user and group)
|
||||
depth: 5 # max depth of object tree storage in FS
|
||||
|
|
|
@ -212,6 +212,8 @@ blobstor:
|
|||
depth: 1
|
||||
width: 4
|
||||
opened_cache_capacity: 50
|
||||
opened_cache_ttl: 5m
|
||||
opened_cache_exp_interval: 15s
|
||||
```
|
||||
|
||||
#### Common options for sub-storages
|
||||
|
@ -228,17 +230,19 @@ blobstor:
|
|||
| `depth` | `int` | `4` | File-system tree depth. |
|
||||
|
||||
#### `blobovnicza` type options
|
||||
| Parameter | Type | Default value | Description |
|
||||
| ----------------------- | ---------- | ------------- | --------------------------------------------------------------------- |
|
||||
| `path` | `string` | | Path to the root of the blobstor. |
|
||||
| `perm` | file mode | `0660` | Default permission for created files and directories. |
|
||||
| `size` | `size` | `1 G` | Maximum size of a single blobovnicza |
|
||||
| `depth` | `int` | `2` | Blobovnicza tree depth. |
|
||||
| `width` | `int` | `16` | Blobovnicza tree width. |
|
||||
| `opened_cache_capacity` | `int` | `16` | Maximum number of simultaneously opened blobovniczas. |
|
||||
| `init_worker_count` | `int` | `5` | Maximum number of concurrent initialization workers. |
|
||||
| `init_in_advance` | `bool` | `false` | If `true`, than all the blobovnicza files will be created on startup. |
|
||||
| `rebuild_drop_timeout` | `duration` | `10s` | Timeout before drop empty blobovnicza file during rebuild. |
|
||||
| Parameter | Type | Default value | Description |
|
||||
|-----------------------------| ---------- |---------------|------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
|
||||
| `path` | `string` | | Path to the root of the blobstor. |
|
||||
| `perm` | file mode | `0660` | Default permission for created files and directories. |
|
||||
| `size` | `size` | `1 G` | Maximum size of a single blobovnicza |
|
||||
| `depth` | `int` | `2` | Blobovnicza tree depth. |
|
||||
| `width` | `int` | `16` | Blobovnicza tree width. |
|
||||
| `opened_cache_capacity` | `int` | `16` | Maximum number of simultaneously opened blobovniczas. |
|
||||
| `opened_cache_ttl` | `duration` | `0` | TTL in cache for opened blobovniczas(disabled by default). In case of heavy random-read and 10 shards each with 10_000 databases and accessing 400 objects per-second we will access each db approximately once per ((10 * 10_000 / 400) = 250 seconds <= 300 seconds = 5 min). Also take in mind that in this scenario they will probably be closed earlier because of the cache capacity, so bigger values are likely to be of no use. |
|
||||
| `opened_cache_exp_interval` | `duration` | `15s` | Cache cleanup interval for expired blobovnicza's. |
|
||||
| `init_worker_count` | `int` | `5` | Maximum number of concurrent initialization workers. |
|
||||
| `init_in_advance` | `bool` | `false` | If `true`, than all the blobovnicza files will be created on startup. |
|
||||
| `rebuild_drop_timeout` | `duration` | `10s` | Timeout before drop empty blobovnicza file during rebuild. |
|
||||
|
||||
### `gc` subsection
|
||||
|
||||
|
|
1
go.mod
1
go.mod
|
@ -16,6 +16,7 @@ require (
|
|||
github.com/chzyer/readline v1.5.1
|
||||
github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc
|
||||
github.com/flynn-archive/go-shlex v0.0.0-20150515145356-3f9db97f8568
|
||||
github.com/go-pkgz/expirable-cache/v3 v3.0.0
|
||||
github.com/google/uuid v1.6.0
|
||||
github.com/hashicorp/golang-lru/v2 v2.0.7
|
||||
github.com/klauspost/compress v1.17.4
|
||||
|
|
2
go.sum
2
go.sum
|
@ -76,6 +76,8 @@ github.com/go-logr/logr v1.4.1 h1:pKouT5E8xu9zeFC39JXRDukb6JFQPXM5p5I91188VAQ=
|
|||
github.com/go-logr/logr v1.4.1/go.mod h1:9T104GzyrTigFIr8wt5mBrctHMim0Nb2HLGrmQ40KvY=
|
||||
github.com/go-logr/stdr v1.2.2 h1:hSWxHoqTgW2S2qGc0LTAI563KZ5YKYRhT3MFKZMbjag=
|
||||
github.com/go-logr/stdr v1.2.2/go.mod h1:mMo/vtBO5dYbehREoey6XUKy/eSumjCCveDpRre4VKE=
|
||||
github.com/go-pkgz/expirable-cache/v3 v3.0.0 h1:u3/gcu3sabLYiTCevoRKv+WzjIn5oo7P8XtiXBeRDLw=
|
||||
github.com/go-pkgz/expirable-cache/v3 v3.0.0/go.mod h1:2OQiDyEGQalYecLWmXprm3maPXeVb5/6/X7yRPYTzec=
|
||||
github.com/go-task/slim-sprig v0.0.0-20210107165309-348f09dbbbc0/go.mod h1:fyg7847qk6SyHyPtNmDHnmrv/HOrqktSC+C9fM+CJOE=
|
||||
github.com/gogo/protobuf v1.3.2/go.mod h1:P1XiOD3dCwIKUDQYPy72D8LYyHL2YPYrpS2s69NZV8Q=
|
||||
github.com/golang/protobuf v1.2.0/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
|
||||
|
|
|
@ -1,6 +1,7 @@
|
|||
package blobovniczatree
|
||||
|
||||
import (
|
||||
"context"
|
||||
"errors"
|
||||
"strconv"
|
||||
"strings"
|
||||
|
@ -72,7 +73,7 @@ const (
|
|||
)
|
||||
|
||||
// NewBlobovniczaTree returns new instance of blobovniczas tree.
|
||||
func NewBlobovniczaTree(opts ...Option) (blz *Blobovniczas) {
|
||||
func NewBlobovniczaTree(ctx context.Context, opts ...Option) (blz *Blobovniczas) {
|
||||
blz = new(Blobovniczas)
|
||||
initConfig(&blz.cfg)
|
||||
|
||||
|
@ -86,7 +87,8 @@ func NewBlobovniczaTree(opts ...Option) (blz *Blobovniczas) {
|
|||
|
||||
blz.commondbManager = newDBManager(blz.rootPath, blz.blzOpts, blz.readOnly, blz.metrics.Blobovnicza(), blz.log)
|
||||
blz.activeDBManager = newActiveDBManager(blz.commondbManager, blz.blzLeafWidth)
|
||||
blz.dbCache = newDBCache(blz.openedCacheSize, blz.commondbManager)
|
||||
blz.dbCache = newDBCache(ctx, blz.openedCacheSize,
|
||||
blz.openedCacheTTL, blz.openedCacheExpInterval, blz.commondbManager)
|
||||
blz.deleteProtectedObjects = newAddressMap()
|
||||
blz.dbFilesGuard = &sync.RWMutex{}
|
||||
blz.rebuildGuard = &sync.RWMutex{}
|
||||
|
|
|
@ -1,11 +1,12 @@
|
|||
package blobovniczatree
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"context"
|
||||
"sync"
|
||||
"time"
|
||||
|
||||
utilSync "git.frostfs.info/TrueCloudLab/frostfs-node/pkg/util/sync"
|
||||
"github.com/hashicorp/golang-lru/v2/simplelru"
|
||||
cache "github.com/go-pkgz/expirable-cache/v3"
|
||||
)
|
||||
|
||||
// dbCache caches sharedDB instances that are NOT open for Put.
|
||||
|
@ -13,30 +14,55 @@ import (
|
|||
// Uses dbManager for opening/closing sharedDB instances.
|
||||
// Stores a reference to an cached sharedDB, so dbManager does not close it.
|
||||
type dbCache struct {
|
||||
cacheGuard *sync.RWMutex
|
||||
cache simplelru.LRUCache[string, *sharedDB]
|
||||
cacheGuard *sync.Mutex
|
||||
cache cache.Cache[string, *sharedDB]
|
||||
pathLock *utilSync.KeyLocker[string] // the order of locks is important: pathLock first, cacheGuard second
|
||||
closed bool
|
||||
nonCached map[string]struct{}
|
||||
wg sync.WaitGroup
|
||||
cancel context.CancelFunc
|
||||
|
||||
dbManager *dbManager
|
||||
}
|
||||
|
||||
func newDBCache(size int, dbManager *dbManager) *dbCache {
|
||||
cache, err := simplelru.NewLRU(size, func(_ string, evictedDB *sharedDB) {
|
||||
evictedDB.Close()
|
||||
})
|
||||
if err != nil {
|
||||
// occurs only if the size is not positive
|
||||
panic(fmt.Errorf("could not create LRU cache of size %d: %w", size, err))
|
||||
}
|
||||
return &dbCache{
|
||||
cacheGuard: &sync.RWMutex{},
|
||||
cache: cache,
|
||||
func newDBCache(parentCtx context.Context, size int,
|
||||
ttl time.Duration, expInterval time.Duration,
|
||||
dbManager *dbManager,
|
||||
) *dbCache {
|
||||
ch := cache.NewCache[string, *sharedDB]().
|
||||
WithTTL(ttl).WithLRU().WithMaxKeys(size).
|
||||
WithOnEvicted(func(_ string, db *sharedDB) {
|
||||
db.Close()
|
||||
})
|
||||
ctx, cancel := context.WithCancel(parentCtx)
|
||||
res := &dbCache{
|
||||
cacheGuard: &sync.Mutex{},
|
||||
wg: sync.WaitGroup{},
|
||||
cancel: cancel,
|
||||
cache: ch,
|
||||
dbManager: dbManager,
|
||||
pathLock: utilSync.NewKeyLocker[string](),
|
||||
nonCached: make(map[string]struct{}),
|
||||
}
|
||||
if ttl > 0 {
|
||||
res.wg.Add(1)
|
||||
go func() {
|
||||
ticker := time.NewTicker(expInterval)
|
||||
defer ticker.Stop()
|
||||
for {
|
||||
select {
|
||||
case <-ctx.Done():
|
||||
res.wg.Done()
|
||||
return
|
||||
case <-ticker.C:
|
||||
res.cacheGuard.Lock()
|
||||
res.cache.DeleteExpired()
|
||||
res.cacheGuard.Unlock()
|
||||
}
|
||||
}
|
||||
}()
|
||||
}
|
||||
return res
|
||||
}
|
||||
|
||||
func (c *dbCache) Open() {
|
||||
|
@ -49,6 +75,8 @@ func (c *dbCache) Open() {
|
|||
func (c *dbCache) Close() {
|
||||
c.cacheGuard.Lock()
|
||||
defer c.cacheGuard.Unlock()
|
||||
c.cancel()
|
||||
c.wg.Wait()
|
||||
c.cache.Purge()
|
||||
c.closed = true
|
||||
}
|
||||
|
@ -88,6 +116,8 @@ func (c *dbCache) getExisted(path string) *sharedDB {
|
|||
|
||||
if value, ok := c.cache.Get(path); ok {
|
||||
return value
|
||||
} else if value != nil {
|
||||
c.cache.Invalidate(path)
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
@ -119,10 +149,9 @@ func (c *dbCache) put(path string, db *sharedDB) bool {
|
|||
|
||||
_, isNonCached := c.nonCached[path]
|
||||
|
||||
if !isNonCached && !c.closed {
|
||||
c.cache.Add(path, db)
|
||||
return true
|
||||
if isNonCached || c.closed {
|
||||
return false
|
||||
}
|
||||
|
||||
return false
|
||||
c.cache.Add(path, db)
|
||||
return true
|
||||
}
|
||||
|
|
|
@ -17,6 +17,7 @@ func TestBlobovniczaTree_Concurrency(t *testing.T) {
|
|||
const n = 1000
|
||||
|
||||
st := NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
WithLogger(test.NewLogger(t)),
|
||||
WithObjectSizeLimit(1024),
|
||||
WithBlobovniczaShallowWidth(10),
|
||||
|
|
|
@ -49,6 +49,7 @@ func createTestTree(t *testing.T, currentDepth, depth, width uint64, path string
|
|||
|
||||
func openAndCloseTestTree(t *testing.T, depth, width uint64, path string) {
|
||||
blz := NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
WithBlobovniczaShallowDepth(depth),
|
||||
WithBlobovniczaShallowWidth(width),
|
||||
WithRootPath(path),
|
||||
|
@ -78,6 +79,7 @@ func TestObjectsAvailableAfterDepthAndWidthEdit(t *testing.T) {
|
|||
rootDir := t.TempDir()
|
||||
|
||||
blz := NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
WithBlobovniczaShallowDepth(3),
|
||||
WithBlobovniczaShallowWidth(5),
|
||||
WithRootPath(rootDir),
|
||||
|
@ -115,6 +117,7 @@ func TestObjectsAvailableAfterDepthAndWidthEdit(t *testing.T) {
|
|||
|
||||
// change depth and width
|
||||
blz = NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
WithBlobovniczaShallowDepth(5),
|
||||
WithBlobovniczaShallowWidth(2),
|
||||
WithRootPath(rootDir),
|
||||
|
@ -152,6 +155,7 @@ func TestObjectsAvailableAfterDepthAndWidthEdit(t *testing.T) {
|
|||
|
||||
// change depth and width back
|
||||
blz = NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
WithBlobovniczaShallowDepth(3),
|
||||
WithBlobovniczaShallowWidth(5),
|
||||
WithRootPath(rootDir),
|
||||
|
|
|
@ -17,6 +17,7 @@ import (
|
|||
func TestExistsInvalidStorageID(t *testing.T) {
|
||||
dir := t.TempDir()
|
||||
b := NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
WithLogger(test.NewLogger(t)),
|
||||
WithObjectSizeLimit(1024),
|
||||
WithBlobovniczaShallowWidth(2),
|
||||
|
|
|
@ -1,6 +1,7 @@
|
|||
package blobovniczatree
|
||||
|
||||
import (
|
||||
"context"
|
||||
"testing"
|
||||
|
||||
"git.frostfs.info/TrueCloudLab/frostfs-node/pkg/local_object_storage/blobstor/common"
|
||||
|
@ -13,6 +14,7 @@ func TestGeneric(t *testing.T) {
|
|||
|
||||
helper := func(t *testing.T, dir string) common.Storage {
|
||||
return NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
WithLogger(test.NewLogger(t)),
|
||||
WithObjectSizeLimit(maxObjectSize),
|
||||
WithBlobovniczaShallowWidth(2),
|
||||
|
@ -40,6 +42,7 @@ func TestControl(t *testing.T) {
|
|||
|
||||
newTree := func(t *testing.T) common.Storage {
|
||||
return NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
WithLogger(test.NewLogger(t)),
|
||||
WithObjectSizeLimit(maxObjectSize),
|
||||
WithBlobovniczaShallowWidth(2),
|
||||
|
|
|
@ -12,6 +12,7 @@ func TestIterateSortedLeavesAndDBPathsAreSame(t *testing.T) {
|
|||
t.Parallel()
|
||||
|
||||
blz := NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
WithBlobovniczaShallowDepth(3),
|
||||
WithBlobovniczaShallowWidth(5),
|
||||
WithRootPath(t.TempDir()),
|
||||
|
|
|
@ -27,32 +27,40 @@ type cfg struct {
|
|||
blzInitWorkerCount int
|
||||
blzMoveBatchSize int
|
||||
createDBInAdvance bool
|
||||
// TTL for blobovnicza's cache
|
||||
openedCacheTTL time.Duration
|
||||
// Interval for deletion expired blobovnicza's
|
||||
openedCacheExpInterval time.Duration
|
||||
}
|
||||
|
||||
type Option func(*cfg)
|
||||
|
||||
const (
|
||||
defaultPerm = 0o700
|
||||
defaultOpenedCacheSize = 50
|
||||
defaultBlzShallowDepth = 2
|
||||
defaultBlzShallowWidth = 16
|
||||
defaultWaitBeforeDropDB = 10 * time.Second
|
||||
defaultBlzInitWorkerCount = 5
|
||||
defaulBlzMoveBatchSize = 10000
|
||||
defaultPerm = 0o700
|
||||
defaultOpenedCacheSize = 50
|
||||
defaultOpenedCacheTTL = 0 // means expiring is off
|
||||
defaultOpenedCacheInterval = 15 * time.Second
|
||||
defaultBlzShallowDepth = 2
|
||||
defaultBlzShallowWidth = 16
|
||||
defaultWaitBeforeDropDB = 10 * time.Second
|
||||
defaultBlzInitWorkerCount = 5
|
||||
defaulBlzMoveBatchSize = 10000
|
||||
)
|
||||
|
||||
func initConfig(c *cfg) {
|
||||
*c = cfg{
|
||||
log: &logger.Logger{Logger: zap.L()},
|
||||
perm: defaultPerm,
|
||||
openedCacheSize: defaultOpenedCacheSize,
|
||||
blzShallowDepth: defaultBlzShallowDepth,
|
||||
blzShallowWidth: defaultBlzShallowWidth,
|
||||
reportError: func(string, error) {},
|
||||
metrics: &noopMetrics{},
|
||||
waitBeforeDropDB: defaultWaitBeforeDropDB,
|
||||
blzInitWorkerCount: defaultBlzInitWorkerCount,
|
||||
blzMoveBatchSize: defaulBlzMoveBatchSize,
|
||||
log: &logger.Logger{Logger: zap.L()},
|
||||
perm: defaultPerm,
|
||||
openedCacheSize: defaultOpenedCacheSize,
|
||||
openedCacheTTL: defaultOpenedCacheTTL,
|
||||
openedCacheExpInterval: defaultOpenedCacheInterval,
|
||||
blzShallowDepth: defaultBlzShallowDepth,
|
||||
blzShallowWidth: defaultBlzShallowWidth,
|
||||
reportError: func(string, error) {},
|
||||
metrics: &noopMetrics{},
|
||||
waitBeforeDropDB: defaultWaitBeforeDropDB,
|
||||
blzInitWorkerCount: defaultBlzInitWorkerCount,
|
||||
blzMoveBatchSize: defaulBlzMoveBatchSize,
|
||||
}
|
||||
}
|
||||
|
||||
|
@ -105,6 +113,18 @@ func WithOpenedCacheSize(sz int) Option {
|
|||
}
|
||||
}
|
||||
|
||||
func WithOpenedCacheTTL(ttl time.Duration) Option {
|
||||
return func(c *cfg) {
|
||||
c.openedCacheTTL = ttl
|
||||
}
|
||||
}
|
||||
|
||||
func WithOpenedCacheExpInterval(expInterval time.Duration) Option {
|
||||
return func(c *cfg) {
|
||||
c.openedCacheExpInterval = expInterval
|
||||
}
|
||||
}
|
||||
|
||||
func WithObjectSizeLimit(sz uint64) Option {
|
||||
return func(c *cfg) {
|
||||
c.blzOpts = append(c.blzOpts, blobovnicza.WithObjectSizeLimit(sz))
|
||||
|
|
|
@ -129,6 +129,7 @@ func testRebuildFailoverObjectDeletedFromSource(t *testing.T) {
|
|||
|
||||
func testRebuildFailoverValidate(t *testing.T, dir string, obj *objectSDK.Object, mustUpdateStorageID bool) {
|
||||
b := NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
WithLogger(test.NewLogger(t)),
|
||||
WithObjectSizeLimit(2048),
|
||||
WithBlobovniczaShallowWidth(2),
|
||||
|
|
|
@ -43,6 +43,7 @@ func TestBlobovniczaTreeRebuildLargeObject(t *testing.T) {
|
|||
|
||||
dir := t.TempDir()
|
||||
b := NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
WithLogger(test.NewLogger(t)),
|
||||
WithObjectSizeLimit(64*1024), // 64KB object size limit
|
||||
WithBlobovniczaShallowWidth(5),
|
||||
|
@ -70,6 +71,7 @@ func TestBlobovniczaTreeRebuildLargeObject(t *testing.T) {
|
|||
require.NoError(t, b.Close())
|
||||
|
||||
b = NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
WithLogger(test.NewLogger(t)),
|
||||
WithObjectSizeLimit(32*1024), // 32KB object size limit
|
||||
WithBlobovniczaShallowWidth(5),
|
||||
|
@ -108,6 +110,7 @@ func TestBlobovniczaTreeRebuildLargeObject(t *testing.T) {
|
|||
func testBlobovniczaTreeRebuildHelper(t *testing.T, sourceDepth, sourceWidth, targetDepth, targetWidth uint64, shouldMigrate bool) {
|
||||
dir := t.TempDir()
|
||||
b := NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
WithLogger(test.NewLogger(t)),
|
||||
WithObjectSizeLimit(2048),
|
||||
WithBlobovniczaShallowWidth(sourceWidth),
|
||||
|
@ -148,6 +151,7 @@ func testBlobovniczaTreeRebuildHelper(t *testing.T, sourceDepth, sourceWidth, ta
|
|||
require.NoError(t, b.Close())
|
||||
|
||||
b = NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
WithLogger(test.NewLogger(t)),
|
||||
WithObjectSizeLimit(2048),
|
||||
WithBlobovniczaShallowWidth(targetWidth),
|
||||
|
|
|
@ -18,6 +18,7 @@ import (
|
|||
|
||||
func defaultTestStorages(p string, smallSizeLimit uint64) ([]SubStorage, *teststore.TestStore, *teststore.TestStore) {
|
||||
smallFileStorage := teststore.New(teststore.WithSubstorage(blobovniczatree.NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
blobovniczatree.WithRootPath(filepath.Join(p, "blobovniczas")),
|
||||
blobovniczatree.WithBlobovniczaShallowWidth(1)), // default width is 16, slow init
|
||||
))
|
||||
|
@ -117,6 +118,7 @@ func TestBlobstor_needsCompression(t *testing.T) {
|
|||
WithStorages([]SubStorage{
|
||||
{
|
||||
Storage: blobovniczatree.NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
blobovniczatree.WithRootPath(filepath.Join(dir, "blobovnicza")),
|
||||
blobovniczatree.WithBlobovniczaShallowWidth(1)), // default width is 16, slow init
|
||||
Policy: func(_ *objectSDK.Object, data []byte) bool {
|
||||
|
|
|
@ -73,6 +73,7 @@ var storages = []storage{
|
|||
desc: "blobovniczatree",
|
||||
create: func(dir string) common.Storage {
|
||||
return blobovniczatree.NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
blobovniczatree.WithRootPath(dir),
|
||||
)
|
||||
},
|
||||
|
|
|
@ -137,6 +137,7 @@ func newStorages(t testing.TB, root string, smallSize uint64) []blobstor.SubStor
|
|||
return []blobstor.SubStorage{
|
||||
{
|
||||
Storage: blobovniczatree.NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
blobovniczatree.WithRootPath(filepath.Join(root, "blobovnicza")),
|
||||
blobovniczatree.WithBlobovniczaShallowDepth(1),
|
||||
blobovniczatree.WithBlobovniczaShallowWidth(1),
|
||||
|
@ -158,6 +159,7 @@ func newStorages(t testing.TB, root string, smallSize uint64) []blobstor.SubStor
|
|||
func newTestStorages(root string, smallSize uint64) ([]blobstor.SubStorage, *teststore.TestStore, *teststore.TestStore) {
|
||||
smallFileStorage := teststore.New(
|
||||
teststore.WithSubstorage(blobovniczatree.NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
blobovniczatree.WithRootPath(filepath.Join(root, "blobovnicza")),
|
||||
blobovniczatree.WithBlobovniczaShallowDepth(1),
|
||||
blobovniczatree.WithBlobovniczaShallowWidth(1),
|
||||
|
|
|
@ -36,6 +36,7 @@ func Test_ObjectNotFoundIfNotDeletedFromMetabase(t *testing.T) {
|
|||
blobstor.WithStorages([]blobstor.SubStorage{
|
||||
{
|
||||
Storage: blobovniczatree.NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
blobovniczatree.WithLogger(test.NewLogger(t)),
|
||||
blobovniczatree.WithRootPath(filepath.Join(rootPath, "blob", "blobovnicza")),
|
||||
blobovniczatree.WithBlobovniczaShallowDepth(1),
|
||||
|
|
|
@ -35,6 +35,7 @@ func TestShard_Lock(t *testing.T) {
|
|||
blobstor.WithStorages([]blobstor.SubStorage{
|
||||
{
|
||||
Storage: blobovniczatree.NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
blobovniczatree.WithRootPath(filepath.Join(rootPath, "blob", "blobovnicza")),
|
||||
blobovniczatree.WithBlobovniczaShallowDepth(2),
|
||||
blobovniczatree.WithBlobovniczaShallowWidth(2)),
|
||||
|
|
|
@ -78,6 +78,7 @@ func testShardGetRange(t *testing.T, hasWriteCache bool) {
|
|||
blobstor.WithStorages([]blobstor.SubStorage{
|
||||
{
|
||||
Storage: blobovniczatree.NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
blobovniczatree.WithLogger(test.NewLogger(t)),
|
||||
blobovniczatree.WithRootPath(filepath.Join(t.TempDir(), "blob", "blobovnicza")),
|
||||
blobovniczatree.WithBlobovniczaShallowDepth(1),
|
||||
|
|
|
@ -59,6 +59,7 @@ func newCustomShard(t testing.TB, enableWriteCache bool, o shardOptions) *Shard
|
|||
blobstor.WithStorages([]blobstor.SubStorage{
|
||||
{
|
||||
Storage: blobovniczatree.NewBlobovniczaTree(
|
||||
context.Background(),
|
||||
blobovniczatree.WithLogger(test.NewLogger(t)),
|
||||
blobovniczatree.WithRootPath(filepath.Join(o.rootPath, "blob", "blobovnicza")),
|
||||
blobovniczatree.WithBlobovniczaShallowDepth(1),
|
||||
|
|
Loading…
Reference in a new issue