moby--moby/vendor/src/github.com/boltdb/bolt

Bolt

Bolt is a pure Go key/value store inspired by Howard Chu's and the LMDB project. The goal of the project is to provide a simple, fast, and reliable database for projects that don't require a full database server such as Postgres or MySQL.

Since Bolt is meant to be used as such a low-level piece of functionality, simplicity is key. The API will be small and only focus on getting values and setting values. That's it.

Project Status

Bolt is stable and the API is fixed. Full unit test coverage and randomized black box testing are used to ensure database consistency and thread safety. Bolt is currently in high-load production environments serving databases as large as 1TB. Many companies such as Shopify and Heroku use Bolt-backed services every day.

Getting Started

Installing

To start using Bolt, install Go and run go get:

$ go get github.com/boltdb/bolt/...

This will retrieve the library and install the bolt command line utility into your $GOBIN path.

Opening a database

The top-level object in Bolt is a DB. It is represented as a single file on your disk and represents a consistent snapshot of your data.

To open your database, simply use the bolt.Open() function:

package main

import (
	"log"

	"github.com/boltdb/bolt"
)

func main() {
	// Open the my.db data file in your current directory.
	// It will be created if it doesn't exist.
	db, err := bolt.Open("my.db", 0600, nil)
	if err != nil {
		log.Fatal(err)
	}
	defer db.Close()

	...
}

Please note that Bolt obtains a file lock on the data file so multiple processes cannot open the same database at the same time. Opening an already open Bolt database will cause it to hang until the other process closes it. To prevent an indefinite wait you can pass a timeout option to the Open() function:

db, err := bolt.Open("my.db", 0600, &bolt.Options{Timeout: 1 * time.Second})

Transactions

Bolt allows only one read-write transaction at a time but allows as many read-only transactions as you want at a time. Each transaction has a consistent view of the data as it existed when the transaction started.

Individual transactions and all objects created from them (e.g. buckets, keys) are not thread safe. To work with data in multiple goroutines you must start a transaction for each one or use locking to ensure only one goroutine accesses a transaction at a time. Creating transaction from the DB is thread safe.

Read-write transactions

To start a read-write transaction, you can use the DB.Update() function:

err := db.Update(func(tx *bolt.Tx) error {
	...
	return nil
})

Inside the closure, you have a consistent view of the database. You commit the transaction by returning nil at the end. You can also rollback the transaction at any point by returning an error. All database operations are allowed inside a read-write transaction.

Always check the return error as it will report any disk failures that can cause your transaction to not complete. If you return an error within your closure it will be passed through.

Read-only transactions

To start a read-only transaction, you can use the DB.View() function:

err := db.View(func(tx *bolt.Tx) error {
	...
	return nil
})

You also get a consistent view of the database within this closure, however, no mutating operations are allowed within a read-only transaction. You can only retrieve buckets, retrieve values, and copy the database within a read-only transaction.

Using buckets

Buckets are collections of key/value pairs within the database. All keys in a bucket must be unique. You can create a bucket using the DB.CreateBucket() function:

db.Update(func(tx *bolt.Tx) error {
	b, err := tx.CreateBucket([]byte("MyBucket"))
	if err != nil {
		return fmt.Errorf("create bucket: %s", err)
	}
	return nil
})

You can also create a bucket only if it doesn't exist by using the Tx.CreateBucketIfNotExists() function. It's a common pattern to call this function for all your top-level buckets after you open your database so you can guarantee that they exist for future transactions.

To delete a bucket, simply call the Tx.DeleteBucket() function.

Using key/value pairs

To save a key/value pair to a bucket, use the Bucket.Put() function:

db.Update(func(tx *bolt.Tx) error {
	b := tx.Bucket([]byte("MyBucket"))
	err := b.Put([]byte("answer"), []byte("42"))
	return err
})

This will set the value of the "answer" key to "42" in the MyBucket bucket. To retrieve this value, we can use the Bucket.Get() function:

db.View(func(tx *bolt.Tx) error {
	b := tx.Bucket([]byte("MyBucket"))
	v := b.Get([]byte("answer"))
	fmt.Printf("The answer is: %s\n", v)
	return nil
})

The Get() function does not return an error because its operation is guarenteed to work (unless there is some kind of system failure). If the key exists then it will return its byte slice value. If it doesn't exist then it will return nil. It's important to note that you can have a zero-length value set to a key which is different than the key not existing.

Use the Bucket.Delete() function to delete a key from the bucket.

Iterating over keys

Bolt stores its keys in byte-sorted order within a bucket. This makes sequential iteration over these keys extremely fast. To iterate over keys we'll use a Cursor:

db.View(func(tx *bolt.Tx) error {
	b := tx.Bucket([]byte("MyBucket"))
	c := b.Cursor()

	for k, v := c.First(); k != nil; k, v = c.Next() {
		fmt.Printf("key=%s, value=%s\n", k, v)
	}

	return nil
})

The cursor allows you to move to a specific point in the list of keys and move forward or backward through the keys one at a time.

The following functions are available on the cursor:

First()  Move to the first key.
Last()   Move to the last key.
Seek()   Move to a specific key.
Next()   Move to the next key.
Prev()   Move to the previous key.

When you have iterated to the end of the cursor then Next() will return nil. You must seek to a position using First(), Last(), or Seek() before calling Next() or Prev(). If you do not seek to a position then these functions will return nil.

Prefix scans

To iterate over a key prefix, you can combine Seek() and bytes.HasPrefix():

db.View(func(tx *bolt.Tx) error {
	c := tx.Bucket([]byte("MyBucket")).Cursor()

	prefix := []byte("1234")
	for k, v := c.Seek(prefix); bytes.HasPrefix(k, prefix); k, v = c.Next() {
		fmt.Printf("key=%s, value=%s\n", k, v)
	}

	return nil
})

Range scans

Another common use case is scanning over a range such as a time range. If you use a sortable time encoding such as RFC3339 then you can query a specific date range like this:

db.View(func(tx *bolt.Tx) error {
	// Assume our events bucket has RFC3339 encoded time keys.
	c := tx.Bucket([]byte("Events")).Cursor()

	// Our time range spans the 90's decade.
	min := []byte("1990-01-01T00:00:00Z")
	max := []byte("2000-01-01T00:00:00Z")

	// Iterate over the 90's.
	for k, v := c.Seek(min); k != nil && bytes.Compare(k, max) != -1; k, v = c.Next() {
		fmt.Printf("%s: %s\n", k, v)
	}

	return nil
})

ForEach()

You can also use the function ForEach() if you know you'll be iterating over all the keys in a bucket:

db.View(func(tx *bolt.Tx) error {
	b := tx.Bucket([]byte("MyBucket"))
	b.ForEach(func(k, v []byte) error {
		fmt.Printf("key=%s, value=%s\n", k, v)
		return nil
	})
	return nil
})

Nested buckets

You can also store a bucket in a key to create nested buckets. The API is the same as the bucket management API on the DB object:

func (*Bucket) CreateBucket(key []byte) (*Bucket, error)
func (*Bucket) CreateBucketIfNotExists(key []byte) (*Bucket, error)
func (*Bucket) DeleteBucket(key []byte) error

Database backups

Bolt is a single file so it's easy to backup. You can use the Tx.Copy() function to write a consistent view of the database to a writer. If you call this from a read-only transaction, it will perform a hot backup and not block your other database reads and writes. It will also use O_DIRECT when available to prevent page cache trashing.

One common use case is to backup over HTTP so you can use tools like cURL to do database backups:

func BackupHandleFunc(w http.ResponseWriter, req *http.Request) {
	err := db.View(func(tx bolt.Tx) error {
		w.Header().Set("Content-Type", "application/octet-stream")
		w.Header().Set("Content-Disposition", `attachment; filename="my.db"`)
		w.Header().Set("Content-Length", strconv.Itoa(int(tx.Size())))
		return tx.Copy(w)
	})
	if err != nil {
		http.Error(w, err.Error(), http.StatusInternalServerError)
	}
}

Then you can backup using this command:

$ curl http://localhost/backup > my.db

Or you can open your browser to http://localhost/backup and it will download automatically.

If you want to backup to another file you can use the Tx.CopyFile() helper function.

Statistics

The database keeps a running count of many of the internal operations it performs so you can better understand what's going on. By grabbing a snapshot of these stats at two points in time we can see what operations were performed in that time range.

For example, we could start a goroutine to log stats every 10 seconds:

go func() {
	// Grab the initial stats.
	prev := db.Stats()

	for {
		// Wait for 10s.
		time.Sleep(10 * time.Second)

		// Grab the current stats and diff them.
		stats := db.Stats()
		diff := stats.Sub(&prev)
		
		// Encode stats to JSON and print to STDERR.
		json.NewEncoder(os.Stderr).Encode(diff)

		// Save stats for the next loop.
		prev = stats
	}
}
}()

It's also useful to pipe these stats to a service such as statsd for monitoring or to provide an HTTP endpoint that will perform a fixed-length sample.

Resources

For more information on getting started with Bolt, check out the following articles:

• Intro to BoltDB: Painless Performant Persistence by Nate Finch.

Comparing Bolt to LMDB

Bolt was originally a port of LMDB so it is architecturally similar. Both use a B+tree, have ACID semanetics with fully serializable transactions, and support lock-free MVCC using a single writer and multiple readers.

The two projects have somewhat diverged. LMDB heavily focuses on raw performance while Bolt has focused on simplicity and ease of use. For example, LMDB allows several unsafe actions such as direct writes and append writes for the sake of performance. Bolt opts to disallow actions which can leave the database in a corrupted state. The only exception to this in Bolt is DB.NoSync.

Caveats & Limitations

It's important to pick the right tool for the job and Bolt is no exception. Here are a few things to note when evaluating and using Bolt:

• Bolt is good for read intensive workloads. Sequential write performance is also fast but random writes can be slow. You can add a write-ahead log or transaction coalescer in front of Bolt to mitigate this issue.

• Bolt uses a B+tree internally so there can be a lot of random page access. SSDs provide a significant performance boost over spinning disks.

• Try to avoid long running read transactions. Bolt uses copy-on-write so old pages cannot be reclaimed while an old transaction is using them.

• Byte slices returned from Bolt are only valid during a transaction. Once the transaction has been committed or rolled back then the memory they point to can be reused by a new page or can be unmapped from virtual memory and you'll see an unexpected fault address panic when accessing it.

• Be careful when using Bucket.FillPercent. Setting a high fill percent for buckets that have random inserts will cause your database to have very poor page utilization.

• Use larger buckets in general. Smaller buckets causes poor page utilization once they become larger than the page size (typically 4KB).

• Bulk loading a lot of random writes into a new bucket can be slow as the page will not split until the transaction is committed. Randomly inserting more than 100,000 key/value pairs into a single new bucket in a single transaction is not advised.

• Bolt uses a memory-mapped file so the underlying operating system handles the caching of the data. Typically, the OS will cache as much of the file as it can in memory and will release memory as needed to other processes. This means that Bolt can show very high memory usage when working with large databases. However, this is expected and the OS will release memory as needed. Bolt can handle databases much larger than the available physical RAM.

Other Projects Using Bolt

Below is a list of public, open source projects that use Bolt:

• Bazil - A file system that lets your data reside where it is most convenient for it to reside.
• DVID - Added Bolt as optional storage engine and testing it against Basho-tuned leveldb.
• Skybox Analytics - A standalone funnel analysis tool for web analytics.
• Scuttlebutt - Uses Bolt to store and process all Twitter mentions of GitHub projects.
• Wiki - A tiny wiki using Goji, BoltDB and Blackfriday.
• ChainStore - Simple key-value interface to a variety of storage engines organized as a chain of operations.
• MetricBase - Single-binary version of Graphite.
• Gitchain - Decentralized, peer-to-peer Git repositories aka "Git meets Bitcoin".
• event-shuttle - A Unix system service to collect and reliably deliver messages to Kafka.
• ipxed - Web interface and api for ipxed.
• BoltStore - Session store using Bolt.
• photosite/session - Sessions for a photo viewing site.
• LedisDB - A high performance NoSQL, using Bolt as optional storage.
• ipLocator - A fast ip-geo-location-server using bolt with bloom filters.
• cayley - Cayley is an open-source graph database using Bolt as optional backend.
• bleve - A pure Go search engine similar to ElasticSearch that uses Bolt as the default storage backend.
• tentacool - REST api server to manage system stuff (IP, DNS, Gateway...) on a linux server.
• SkyDB - Behavioral analytics database.

If you are using Bolt in a project please send a pull request to add it to the list.