2019-11-29 01:11:52 -07:00
|
|
|
// Copyright (C) 2019 The Syncthing Authors.
|
|
|
|
//
|
|
|
|
// This Source Code Form is subject to the terms of the Mozilla Public
|
|
|
|
// License, v. 2.0. If a copy of the MPL was not distributed with this file,
|
|
|
|
// You can obtain one at https://mozilla.org/MPL/2.0/.
|
|
|
|
|
|
|
|
package backend
|
|
|
|
|
|
|
|
import (
|
|
|
|
"sync"
|
|
|
|
)
|
|
|
|
|
|
|
|
// The Reader interface specifies the read-only operations available on the
|
|
|
|
// main database and on read-only transactions (snapshots). Note that when
|
|
|
|
// called directly on the database handle these operations may take implicit
|
|
|
|
// transactions and performance may suffer.
|
|
|
|
type Reader interface {
|
|
|
|
Get(key []byte) ([]byte, error)
|
|
|
|
NewPrefixIterator(prefix []byte) (Iterator, error)
|
|
|
|
NewRangeIterator(first, last []byte) (Iterator, error)
|
|
|
|
}
|
|
|
|
|
|
|
|
// The Writer interface specifies the mutating operations available on the
|
|
|
|
// main database and on writable transactions. Note that when called
|
|
|
|
// directly on the database handle these operations may take implicit
|
|
|
|
// transactions and performance may suffer.
|
|
|
|
type Writer interface {
|
|
|
|
Put(key, val []byte) error
|
|
|
|
Delete(key []byte) error
|
|
|
|
}
|
|
|
|
|
|
|
|
// The ReadTransaction interface specifies the operations on read-only
|
|
|
|
// transactions. Every ReadTransaction must be released when no longer
|
|
|
|
// required.
|
|
|
|
type ReadTransaction interface {
|
|
|
|
Reader
|
|
|
|
Release()
|
|
|
|
}
|
|
|
|
|
|
|
|
// The WriteTransaction interface specifies the operations on writable
|
|
|
|
// transactions. Every WriteTransaction must be either committed or released
|
|
|
|
// (i.e., discarded) when no longer required. No further operations must be
|
|
|
|
// performed after release or commit (regardless of whether commit succeeded),
|
|
|
|
// with one exception -- it's fine to release an already committed or released
|
|
|
|
// transaction.
|
|
|
|
//
|
|
|
|
// A Checkpoint is a potential partial commit of the transaction so far, for
|
|
|
|
// purposes of saving memory when transactions are in-RAM. Note that
|
|
|
|
// transactions may be checkpointed *anyway* even if this is not called, due to
|
|
|
|
// resource constraints, but this gives you a chance to decide when.
|
2020-02-13 07:23:08 -07:00
|
|
|
//
|
|
|
|
// Functions can be passed to Checkpoint. These are run if and only if the
|
|
|
|
// checkpoint will result in a flush, and will run before the flush. The
|
|
|
|
// transaction can be accessed via a closure. If an error is returned from
|
|
|
|
// these functions the flush will be aborted and the error bubbled.
|
2019-11-29 01:11:52 -07:00
|
|
|
type WriteTransaction interface {
|
|
|
|
ReadTransaction
|
|
|
|
Writer
|
2020-02-13 07:23:08 -07:00
|
|
|
Checkpoint(...func() error) error
|
2019-11-29 01:11:52 -07:00
|
|
|
Commit() error
|
|
|
|
}
|
|
|
|
|
|
|
|
// The Iterator interface specifies the operations available on iterators
|
|
|
|
// returned by NewPrefixIterator and NewRangeIterator. The iterator pattern
|
|
|
|
// is to loop while Next returns true, then check Error after the loop. Next
|
|
|
|
// will return false when iteration is complete (Error() == nil) or when
|
|
|
|
// there is an error preventing iteration, which is then returned by
|
|
|
|
// Error(). For example:
|
|
|
|
//
|
|
|
|
// it, err := db.NewPrefixIterator(nil)
|
|
|
|
// if err != nil {
|
|
|
|
// // problem preventing iteration
|
|
|
|
// }
|
|
|
|
// defer it.Release()
|
|
|
|
// for it.Next() {
|
|
|
|
// // ...
|
|
|
|
// }
|
|
|
|
// if err := it.Error(); err != nil {
|
|
|
|
// // there was a database problem while iterating
|
|
|
|
// }
|
|
|
|
//
|
|
|
|
// An iterator must be Released when no longer required. The Error method
|
|
|
|
// can be called either before or after Release with the same results. If an
|
|
|
|
// iterator was created in a transaction (whether read-only or write) it
|
|
|
|
// must be released before the transaction is released (or committed).
|
|
|
|
type Iterator interface {
|
|
|
|
Next() bool
|
|
|
|
Key() []byte
|
|
|
|
Value() []byte
|
|
|
|
Error() error
|
|
|
|
Release()
|
|
|
|
}
|
|
|
|
|
|
|
|
// The Backend interface represents the main database handle. It supports
|
|
|
|
// both read/write operations and opening read-only or writable
|
|
|
|
// transactions. Depending on the actual implementation, individual
|
|
|
|
// read/write operations may be implicitly wrapped in transactions, making
|
|
|
|
// them perform quite badly when used repeatedly. For bulk operations,
|
|
|
|
// consider always using a transaction of the appropriate type. The
|
|
|
|
// transaction isolation level is "read committed" - there are no dirty
|
|
|
|
// reads.
|
|
|
|
type Backend interface {
|
|
|
|
Reader
|
|
|
|
Writer
|
|
|
|
NewReadTransaction() (ReadTransaction, error)
|
|
|
|
NewWriteTransaction() (WriteTransaction, error)
|
|
|
|
Close() error
|
lib/db: Deduplicate block lists in database (fixes #5898) (#6283)
* lib/db: Deduplicate block lists in database (fixes #5898)
This moves the block list in the database out from being just a field on
the FileInfo to being an object of its own. When putting a FileInfo we
marshal the block list separately and store it keyed by the sha256 of
the marshalled block list. When getting, if we are not doing a
"truncated" get, we do an extra read and unmarshal for the block list.
Old block lists are cleared out by a periodic GC sweep. The alternative
would be to use refcounting, but:
- There is a larger risk of getting that wrong and either dropping a
block list in error or keeping them around forever.
- It's tricky with our current database, as we don't have dirty reads.
This means that if we update two FileInfos with identical block lists in
the same transaction we can't just do read/modify/write for the ref
counters as we wouldn't see our own first update. See above about
tracking this and risks about getting it wrong.
GC uses a bloom filter for keys to avoid heavy RAM usage. GC can't run
concurrently with FileInfo updates so there is a new lock around those
operation at the lowlevel.
The end result is a much more compact database, especially for setups
with many peers where files get duplicated many times.
This is per-key-class stats for a large database I'm currently working
with, under the current schema:
```
0x00: 9138161 items, 870876 KB keys + 7397482 KB data, 95 B + 809 B avg, 1637651 B max
0x01: 185656 items, 10388 KB keys + 1790909 KB data, 55 B + 9646 B avg, 924525 B max
0x02: 916890 items, 84795 KB keys + 3667 KB data, 92 B + 4 B avg, 192 B max
0x03: 384 items, 27 KB keys + 5 KB data, 72 B + 15 B avg, 87 B max
0x04: 1109 items, 17 KB keys + 17 KB data, 15 B + 15 B avg, 69 B max
0x06: 383 items, 3 KB keys + 0 KB data, 9 B + 2 B avg, 18 B max
0x07: 510 items, 4 KB keys + 12 KB data, 9 B + 24 B avg, 41 B max
0x08: 1349 items, 12 KB keys + 10 KB data, 9 B + 8 B avg, 17 B max
0x09: 194 items, 0 KB keys + 123 KB data, 5 B + 634 B avg, 11484 B max
0x0a: 3 items, 0 KB keys + 0 KB data, 14 B + 7 B avg, 30 B max
0x0b: 181836 items, 2363 KB keys + 10694 KB data, 13 B + 58 B avg, 173 B max
Total 10426475 items, 968490 KB keys + 9202925 KB data.
```
Note 7.4 GB of data in class 00, total size 9.2 GB. After running the
migration we get this instead:
```
0x00: 9138161 items, 870876 KB keys + 2611392 KB data, 95 B + 285 B avg, 4788 B max
0x01: 185656 items, 10388 KB keys + 1790909 KB data, 55 B + 9646 B avg, 924525 B max
0x02: 916890 items, 84795 KB keys + 3667 KB data, 92 B + 4 B avg, 192 B max
0x03: 384 items, 27 KB keys + 5 KB data, 72 B + 15 B avg, 87 B max
0x04: 1109 items, 17 KB keys + 17 KB data, 15 B + 15 B avg, 69 B max
0x06: 383 items, 3 KB keys + 0 KB data, 9 B + 2 B avg, 18 B max
0x07: 510 items, 4 KB keys + 12 KB data, 9 B + 24 B avg, 41 B max
0x09: 194 items, 0 KB keys + 123 KB data, 5 B + 634 B avg, 11484 B max
0x0a: 3 items, 0 KB keys + 0 KB data, 14 B + 17 B avg, 51 B max
0x0b: 181836 items, 2363 KB keys + 10694 KB data, 13 B + 58 B avg, 173 B max
0x0d: 44282 items, 1461 KB keys + 61081 KB data, 33 B + 1379 B avg, 1637399 B max
Total 10469408 items, 969939 KB keys + 4477905 KB data.
```
Class 00 is now down to 2.6 GB, with just 61 MB added in class 0d.
There will be some additional reads in some cases which theoretically
hurts performance, but this will be more than compensated for by smaller
writes and better compaction.
On my own home setup which just has three devices and a handful of
folders the difference is smaller in absolute numbers of course, but
still less than half the old size:
```
0x00: 297122 items, 20894 KB keys + 306860 KB data, 70 B + 1032 B avg, 103237 B max
0x01: 115299 items, 7738 KB keys + 17542 KB data, 67 B + 152 B avg, 419 B max
0x02: 1430537 items, 121223 KB keys + 5722 KB data, 84 B + 4 B avg, 253 B max
...
Total 1947412 items, 151268 KB keys + 337485 KB data.
```
to:
```
0x00: 297122 items, 20894 KB keys + 37038 KB data, 70 B + 124 B avg, 520 B max
0x01: 115299 items, 7738 KB keys + 17542 KB data, 67 B + 152 B avg, 419 B max
0x02: 1430537 items, 121223 KB keys + 5722 KB data, 84 B + 4 B avg, 253 B max
...
0x0d: 18041 items, 595 KB keys + 71964 KB data, 33 B + 3988 B avg, 101109 B max
Total 1965447 items, 151863 KB keys + 139628 KB data.
```
* wip
* wip
* wip
* wip
2020-01-24 00:35:44 -07:00
|
|
|
Compact() error
|
2019-11-29 01:11:52 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
type Tuning int
|
|
|
|
|
|
|
|
const (
|
|
|
|
// N.b. these constants must match those in lib/config.Tuning!
|
|
|
|
TuningAuto Tuning = iota
|
|
|
|
TuningSmall
|
|
|
|
TuningLarge
|
|
|
|
)
|
|
|
|
|
|
|
|
func Open(path string, tuning Tuning) (Backend, error) {
|
|
|
|
return OpenLevelDB(path, tuning)
|
|
|
|
}
|
|
|
|
|
|
|
|
func OpenMemory() Backend {
|
|
|
|
return OpenLevelDBMemory()
|
|
|
|
}
|
|
|
|
|
|
|
|
type errClosed struct{}
|
|
|
|
|
|
|
|
func (errClosed) Error() string { return "database is closed" }
|
|
|
|
|
|
|
|
type errNotFound struct{}
|
|
|
|
|
|
|
|
func (errNotFound) Error() string { return "key not found" }
|
|
|
|
|
|
|
|
func IsClosed(err error) bool {
|
|
|
|
if _, ok := err.(errClosed); ok {
|
|
|
|
return true
|
|
|
|
}
|
|
|
|
if _, ok := err.(*errClosed); ok {
|
|
|
|
return true
|
|
|
|
}
|
|
|
|
return false
|
|
|
|
}
|
|
|
|
|
|
|
|
func IsNotFound(err error) bool {
|
|
|
|
if _, ok := err.(errNotFound); ok {
|
|
|
|
return true
|
|
|
|
}
|
|
|
|
if _, ok := err.(*errNotFound); ok {
|
|
|
|
return true
|
|
|
|
}
|
|
|
|
return false
|
|
|
|
}
|
|
|
|
|
|
|
|
// releaser manages counting on top of a waitgroup
|
|
|
|
type releaser struct {
|
2020-02-11 06:31:43 -07:00
|
|
|
wg *closeWaitGroup
|
2019-11-29 01:11:52 -07:00
|
|
|
once *sync.Once
|
|
|
|
}
|
|
|
|
|
2020-02-11 06:31:43 -07:00
|
|
|
func newReleaser(wg *closeWaitGroup) (*releaser, error) {
|
|
|
|
if err := wg.Add(1); err != nil {
|
|
|
|
return nil, err
|
|
|
|
}
|
2019-11-29 01:11:52 -07:00
|
|
|
return &releaser{
|
|
|
|
wg: wg,
|
|
|
|
once: new(sync.Once),
|
2020-02-11 06:31:43 -07:00
|
|
|
}, nil
|
2019-11-29 01:11:52 -07:00
|
|
|
}
|
|
|
|
|
|
|
|
func (r releaser) Release() {
|
|
|
|
// We use the Once because we may get called multiple times from
|
|
|
|
// Commit() and deferred Release().
|
|
|
|
r.once.Do(func() {
|
|
|
|
r.wg.Done()
|
|
|
|
})
|
|
|
|
}
|
2020-02-11 06:31:43 -07:00
|
|
|
|
|
|
|
// closeWaitGroup behaves just like a sync.WaitGroup, but does not require
|
|
|
|
// a single routine to do the Add and Wait calls. If Add is called after
|
|
|
|
// CloseWait, it will return an error, and both are safe to be used concurrently.
|
|
|
|
type closeWaitGroup struct {
|
|
|
|
sync.WaitGroup
|
|
|
|
closed bool
|
|
|
|
closeMut sync.RWMutex
|
|
|
|
}
|
|
|
|
|
|
|
|
func (cg *closeWaitGroup) Add(i int) error {
|
|
|
|
cg.closeMut.RLock()
|
|
|
|
defer cg.closeMut.RUnlock()
|
|
|
|
if cg.closed {
|
|
|
|
return errClosed{}
|
|
|
|
}
|
|
|
|
cg.WaitGroup.Add(i)
|
|
|
|
return nil
|
|
|
|
}
|
|
|
|
|
|
|
|
func (cg *closeWaitGroup) CloseWait() {
|
|
|
|
cg.closeMut.Lock()
|
|
|
|
cg.closed = true
|
|
|
|
cg.closeMut.Unlock()
|
|
|
|
cg.WaitGroup.Wait()
|
|
|
|
}
|