Copy From bson.go(gopkg.in/mgo.v2 v2.0.0-20190816093944-a6b53ec6cb22)

BarrierMatrix only split oplogs into segments. without convert stage stage.

Batcher is used to batch oplog before sending in order to improve performance.

BulkWriter use general bulk interface such like Insert/Update/Delete to execute command

* * ====== ChecksumCoder ======= * */

CommandWriter use run_command to execute command

* * ====== Compressor ======= * */

Cursor

DBSyncer (1 shard -> 1 DBSyncer)

DocCollections is for result of config.collections

DocFilter: AutologousFilter, NamespaceFilter

DocumentReader is the reader for single piece

DocumentSplitter pre-split the big collection into several pieces

Event example: * { _id : { // 存储元信息 "_data" : <BinData|hex string> // resumeToken }, "operationType" : "<operation>", // insert, delete, replace,

** * File Structure * * |----- Header ------|------ OplogBlock ------|------ OplogBlock --------| ...... * |<--- 32bytes ---->| * */

HttpApiCheckpoint for http api

below structs are for decode result of listIndexes IndexInfo

queue size info

ListIndexesResult

struct used to mark the metric delta. Value: current value Delta: the difference between current value and previous value previous: store the previous

MongoCheckpoint for mongo

************************************************/

NamespaceFilter because regexp use the default perl engine which is not support inverse match, so use two rules to match

OpTypeFilter filters oplogs by oplog op type instead of only command oplogs.

OplogFilter include: AutologousFilter, NamespaceFilter, GidFilter, NoopFilter, DDLFilter

OplogReader represents stream reader from mongodb that specified by an url. And with query options. user can iterate oplogs.

OplogSyncer poll oplogs from original source MongoDB.