Skip to content

Commit

Permalink
services: add new service for fetching blocks from NeoFS
Browse files Browse the repository at this point in the history
Close #3496

Co-authored-by: Anna Shaleva <[email protected]>
Signed-off-by: Ekaterina Pavlova <[email protected]>
  • Loading branch information
AliceInHunterland and AnnaShaleva committed Sep 10, 2024
1 parent 5dff3fc commit d884ae9
Show file tree
Hide file tree
Showing 10 changed files with 871 additions and 47 deletions.
13 changes: 13 additions & 0 deletions config/protocol.testnet.yml
Original file line number Diff line number Diff line change
Expand Up @@ -100,3 +100,16 @@ ApplicationConfiguration:
Enabled: false
Addresses:
- ":2113"
NeoFSBlockFetcher:
Enabled: false
Addresses:
- st1.storage.fs.neo.org:8080
Timeout: 10m
DownloaderWorkersCount: 500
OIDBatchSize: 8000
BQueueSize: 16000 # must be larger than OIDBatchSize; recommended to be 2*OIDBatchSize or 3*OIDBatchSize
SkipIndexFilesSearch: false
IndexFileSize: 128000
ContainerID: "EPGuD26wYgQJbmDdVBoYoNZiMKHwFMJT3A5WqPjdUHxH"
BlockAttribute: "block"
IndexFileAttribute: "oid"
74 changes: 74 additions & 0 deletions docs/neofs-blockstorage.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,74 @@
# NeoFS block storage

Using NeoFS to store chain's blocks and snapshots was proposed in
[#3463](https://github.com/neo-project/neo/issues/3463). NeoGo contains several
extensions utilizing NeoFS block storage aimed to improve node synchronization
efficiency and reduce node storage size.

## Components and functionality

### Block storage schema

A single NeoFS container is used to store blocks and index files. Each block
is stored in a binary form as a separate object with a unique OID and a set of
attributes:
- block object identifier with block index value (`block:1`)
- primary node index (`primary:0`)
- block hash in the LE form (`hash:5412a781caf278c0736556c0e544c7cfdbb6e3c62ae221ef53646be89364566b`)
- previous block hash in the LE form (`prevHash:3654a054d82a8178c7dfacecc2c57282e23468a42ee407f14506368afe22d929`)
- millisecond-precision block timestamp (`time:1627894840919`)

Each index file is an object containing a constant-sized batch of raw block object
IDs in binary form ordered by block index. Each index file is marked with the
following attributes:
- index file identifier with consecutive file index value (`oid:0`)
- the number of OIDs included into index file (`size:128000`)

### NeoFS BlockFetcher

NeoFS BlockFetcher service is designed as an alternative to P2P synchronisation
protocol. It allows to download blocks from a trusted container in the NeoFS network
and persist them to database using standard verification flow. NeoFS BlockFetcher
service primarily used during the node's bootstrap, providing a fast alternative to
P2P blocks synchronisation.

NeoFS BlockFetcher service has two modes of operation:
- Index File Search: Search for index files, which contain batches of block object
IDs and fetch blocks from NeoFS by retrieved OIDs.
- Direct Block Search: Search and fetch blocks directly from NeoFS container via
built-in NeoFS object search mechanism.

Operation mode of BlockFetcher can be configured via `SkipIndexFilesSearch`
parameter.

#### Operation flow

1. **OID Fetching**:
Depending on the mode, the service either:
- Searches for index files by index file attribute and reads block OIDs from index
file object-by-object.
- Searches batches of blocks directly by block attribute (the batch size is
configured via `OIDBatchSize` parameter).

Once the OIDs are retrieved, they are immediately redirected to the
block downloading routines for further processing. The channel that
is used to redirect block OIDs to downloading routines is buffered
to provide smooth OIDs delivery without delays. The size of this channel
can be configured via `OIDBatchSize` parameter and equals to `2*OIDBatchSize`.
2. **Parallel Block Downloading**:
The number of downloading routines can be configured via
`DownloaderWorkersCount` parameter. It's up to the user to find the
balance between the downloading speed and blocks persist speed for every
node that uses NeoFS BlockFetcher. Downloaded blocks are placed into a
buffered channel of size `IDBatchSize` with further redirection to the
block queue.
3. **Block Insertion**:
Downloaded blocks are inserted into the blockchain using the same logic
as in the P2P synchronisation protocol. The block queue is used to order
downloaded blocks before they are inserted into the blockchain. The
size of the queue can be configured via the `BQueueSize` parameter
and should be larger than the `OIDBatchSize` parameter to avoid blocking
the downloading routines.

Once all blocks available in the NeoFS container are processed, the service
shuts down automatically.
50 changes: 50 additions & 0 deletions docs/node-configuration.md
Original file line number Diff line number Diff line change
Expand Up @@ -21,6 +21,7 @@ node-related settings described in the table below.
| GarbageCollectionPeriod | `uint32` | 10000 | Controls MPT garbage collection interval (in blocks) for configurations with `RemoveUntraceableBlocks` enabled and `KeepOnlyLatestState` disabled. In this mode the node stores a number of MPT trees (corresponding to `MaxTraceableBlocks` and `StateSyncInterval`), but the DB needs to be clean from old entries from time to time. Doing it too often will cause too much processing overhead, doing it too rarely will leave more useless data in the DB. |
| KeepOnlyLatestState | `bool` | `false` | Specifies if MPT should only store the latest state (or a set of latest states, see `P2PStateExchangeExtensions` section in the ProtocolConfiguration for details). If true, DB size will be smaller, but older roots won't be accessible. This value should remain the same for the same database. | |
| LogPath | `string` | "", so only console logging | File path where to store node logs. |
| NeoFSBlockFetcher | [NeoFS BlockFetcher Configuration](#NeoFS-BlockFetcher-Configuration) | | NeoFS BlockFetcher module configuration. See the [NeoFS BlockFetcher Configuration](#NeoFS-BlockFetcher-Configuration) section for details. |
| Oracle | [Oracle Configuration](#Oracle-Configuration) | | Oracle module configuration. See the [Oracle Configuration](#Oracle-Configuration) section for details. |
| P2P | [P2P Configuration](#P2P-Configuration) | | Configuration values for P2P network interaction. See the [P2P Configuration](#P2P-Configuration) section for details. |
| P2PNotary | [P2P Notary Configuration](#P2P-Notary-Configuration) | | P2P Notary module configuration. See the [P2P Notary Configuration](#P2P-Notary-Configuration) section for details. |
Expand Down Expand Up @@ -153,6 +154,55 @@ where:
Please, refer to the [Notary module documentation](./notary.md#Notary node module) for
details on module features.

### NeoFS BlockFetcher Configuration

`NeoFSBlockFetcher` configuration section contains settings for NeoFS
BlockFetcher module and has the following structure:
```
NeoFSBlockFetcher:
Enabled: true
UnlockWallet:
Path: "./wallet.json"
Password: "pass"
Addresses:
- st1.storage.fs.neo.org:8080
Timeout: 10m
DownloaderWorkersCount: 500
OIDBatchSize: 8000
BQueueSize: 16000
SkipIndexFilesSearch: false
ContainerID: "EPGuD26wYgQJbmDdVBoYoNZiMKHwFMJT3A5WqPjdUHxH"
BlockAttribute: "block"
IndexFileAttribute: "oid"
IndexFileSize: 128000
```
where:
- `Enabled` enables NeoFS BlockFetcher module.
- `UnlockWallet` contains wallet settings to retrieve account to sign requests to
NeoFS. Without this setting, the module will use randomly generated private key.
For configuration details see [Unlock Wallet Configuration](#Unlock-Wallet-Configuration)
- `Addresses` is a list of NeoFS storage nodes addresses.
- `Timeout` is a timeout for a single request to NeoFS storage node.
- `ContainerID` is a container ID to fetch blocks from.
- `BlockAttribute` is an attribute name of NeoFS object that contains block
data.
- `IndexFileAttribute` is an attribute name of NeoFS index object that contains block
object IDs.
- `DownloaderWorkersCount` is a number of workers that download blocks from
NeoFS in parallel.
- `OIDBatchSize` is the number of blocks to search per a single request to NeoFS
in case of disabled index files search. Also, for both modes of BlockFetcher
operation this setting manages the buffer size of OIDs and blocks transferring
channels.
- `BQueueSize` is a size of the block queue used to manage consecutive blocks
addition to the chain. It must be larger than `OIDBatchSize` and highly recommended
to be `2*OIDBatchSize` or `3*OIDBatchSize`.
- `SkipIndexFilesSearch` is a flag that allows to skip index files search and search
for blocks directly. It is set to `false` by default.
- `IndexFileSize` is the number of OID objects stored in the index files. This
setting depends on the NeoFS block storage configuration and is applicable only if
`SkipIndexFilesSearch` is set to `false`.

### Metrics Services Configuration

Metrics services configuration describes options for metrics services (pprof,
Expand Down
23 changes: 17 additions & 6 deletions pkg/config/application_config.go
Original file line number Diff line number Diff line change
Expand Up @@ -23,12 +23,13 @@ type ApplicationConfiguration struct {
Pprof BasicService `yaml:"Pprof"`
Prometheus BasicService `yaml:"Prometheus"`

Relay bool `yaml:"Relay"`
Consensus Consensus `yaml:"Consensus"`
RPC RPC `yaml:"RPC"`
Oracle OracleConfiguration `yaml:"Oracle"`
P2PNotary P2PNotary `yaml:"P2PNotary"`
StateRoot StateRoot `yaml:"StateRoot"`
Relay bool `yaml:"Relay"`
Consensus Consensus `yaml:"Consensus"`
RPC RPC `yaml:"RPC"`
Oracle OracleConfiguration `yaml:"Oracle"`
P2PNotary P2PNotary `yaml:"P2PNotary"`
StateRoot StateRoot `yaml:"StateRoot"`
NeoFSBlockFetcher NeoFSBlockFetcher `yaml:"NeoFSBlockFetcher"`
}

// EqualsButServices returns true when the o is the same as a except for services
Expand Down Expand Up @@ -141,3 +142,13 @@ func (a *ApplicationConfiguration) GetAddresses() ([]AnnounceableAddress, error)
}
return addrs, nil
}

// Validate checks ApplicationConfiguration for internal consistency and returns
// an error if any invalid settings are found. This ensures that the application
// configuration is valid and safe to use for further operations.
func (a *ApplicationConfiguration) Validate() error {
if err := a.NeoFSBlockFetcher.Validate(); err != nil {
return fmt.Errorf("invalid NeoFSBlockFetcher config: %w", err)
}
return nil
}
51 changes: 51 additions & 0 deletions pkg/config/blockfetcher_config.go
Original file line number Diff line number Diff line change
@@ -0,0 +1,51 @@
package config

import (
"errors"
"fmt"
"time"

cid "github.com/nspcc-dev/neofs-sdk-go/container/id"
)

// NeoFSBlockFetcher represents the configuration for the NeoFS BlockFetcher service.
type NeoFSBlockFetcher struct {
InternalService `yaml:",inline"`
Timeout time.Duration `yaml:"Timeout"`
ContainerID string `yaml:"ContainerID"`
Addresses []string `yaml:"Addresses"`
OIDBatchSize int `yaml:"OIDBatchSize"`
BlockAttribute string `yaml:"BlockAttribute"`
IndexFileAttribute string `yaml:"IndexFileAttribute"`
DownloaderWorkersCount int `yaml:"DownloaderWorkersCount"`
BQueueSize int `yaml:"BQueueSize"`
SkipIndexFilesSearch bool `yaml:"SkipIndexFilesSearch"`
IndexFileSize uint32 `yaml:"IndexFileSize"`
}

// Validate checks NeoFSBlockFetcher for internal consistency and ensures
// that all required fields are properly set. It returns an error if the
// configuration is invalid or if the ContainerID cannot be properly decoded.
func (cfg *NeoFSBlockFetcher) Validate() error {
if !cfg.Enabled {
return nil
}
if cfg.ContainerID == "" {
return errors.New("container ID is not set")
}
var containerID cid.ID
err := containerID.DecodeString(cfg.ContainerID)
if err != nil {
return fmt.Errorf("invalid container ID: %w", err)
}
if cfg.BQueueSize < cfg.OIDBatchSize {
return fmt.Errorf("BQueueSize (%d) is lower than OIDBatchSize (%d)", cfg.BQueueSize, cfg.OIDBatchSize)
}
if len(cfg.Addresses) == 0 {
return errors.New("addresses are not set")
}
if !cfg.SkipIndexFilesSearch && cfg.IndexFileSize == 0 {
return errors.New("IndexFileSize is not set")
}
return nil
}
4 changes: 4 additions & 0 deletions pkg/config/config.go
Original file line number Diff line number Diff line change
Expand Up @@ -116,6 +116,10 @@ func LoadFile(configPath string, relativePath ...string) (Config, error) {
if err != nil {
return Config{}, err
}
err = config.ApplicationConfiguration.Validate()
if err != nil {
return Config{}, err
}

return config, nil
}
Expand Down
Loading

0 comments on commit d884ae9

Please sign in to comment.