MongoDB. local Houston 2019: Distributed Transactions: With Great Power Comes Great Responsibility [MongoDB]
1 like309 views
The document discusses MongoDB transactions, highlighting their ACID compliance and ease of implementation across shards. It covers transaction guidelines, like retaining existing data modeling rules, the necessity of retry logic, and cautions against excessive use of transactions. Additionally, it presents new functionalities in MongoDB 4.2, such as larger transaction sizes and improved performance metrics, while emphasizing best practices for optimizing transaction execution.
![Read Your Own Writes
Operation
Read
TXN
{_id: “OWNER12345”,
name: { first: “Aly”,
last: “Cabral”},
accountType: “Paid”,
pets: [ { type: “Dog”,
name: “Lady”}],
contact: { phoneNumber: 8008675309},
Address: { city: “New York”,
street: “94th”,
zipecode: 10025}}
Uncommitted Writes
street: “92nd”
Address.street = 94th
Address.street = 92nd
@aly_cabral](https://image.slidesharecdn.com/mongodb4-190820132749/85/MongoDB-local-Houston-2019-Distributed-Transactions-With-Great-Power-Comes-Great-Responsibility-MongoDB-10-320.jpg)
![MongoDB. local Houston 2019: Distributed Transactions: With Great Power Comes Great Responsibility [MongoDB]](https://image.slidesharecdn.com/mongodb4-190820132749/85/MongoDB-local-Houston-2019-Distributed-Transactions-With-Great-Power-Comes-Great-Responsibility-MongoDB-36-320.jpg)
MongoDB. local Houston 2019: Distributed Transactions: With Great Power Comes Great Responsibility [MongoDB]
- 1. Distributed Transactions: with Great Power Comes Great Responsibility Alyson Cabral, Product Manager @aly_cabral
- 2. I know what you’re thinking… @aly_cabral
- 3. But, Really… MongoDB Transactions are ACID Compliant (yes, even across shards) @aly_cabral
- 4. Transactions make application development easier 1. Don’t need to worry about interim states 2. Rollback/Abort is handled automatically 3. Point-in-time consistency @aly_cabral
- 5. def insertDocuments(s): s.client.db.col.insert_one({'abc': 1}, session=s) s.client.db.col.insert_one({'xyz': 999}, session=s) with client.start_session() as s: s.with_transaction(insertDocuments); New API in MongoDB 4.2! @aly_cabral
- 8. All or Nothing Execution TXN Application MongoDB Uncommitted Writes { Document 1 } { Document 2 } { Document 3 } Committed Writes @aly_cabral
- 9. Snapshot Isolation 100 101 102 103 104 105 106 107 108 109 110 111 Snapshot 102 Snapshot 108 @aly_cabral
- 10. Read Your Own Writes Operation Read TXN {_id: “OWNER12345”, name: { first: “Aly”, last: “Cabral”}, accountType: “Paid”, pets: [ { type: “Dog”, name: “Lady”}], contact: { phoneNumber: 8008675309}, Address: { city: “New York”, street: “94th”, zipecode: 10025}} Uncommitted Writes street: “92nd” Address.street = 94th Address.street = 92nd @aly_cabral
- 12. 1) All data modeling rules still apply 2) Transactions should not be the most common operation (sanity check for rule 1) 3) Pass in the session to all statements 4) Implement retry logic, transactions can always abort 5) Don’t unnecessarily leave snapshots open 6) To trigger write conflicts, make sure you’re doing writes 7) Plan for DDL Operations @aly_cabral
- 15. Participant any shard that executes operations on behalf of a given transaction Coordinator A single shard that is responsible for coordinating the commit across shards for a given transaction @aly_cabral
- 16. P S S S P S Query Router Insert {_id : “abc”}Insert {_id : “xyz”} Coordinator! Commit transaction @aly_cabral
- 17. “ We see at least 2x performance improvement when running MongoDB transactions vs logic on our own, in- app” Imagine a world without transactions @aly_cabral
- 18. For even better perf results… collocate the data @aly_cabral
- 19. If you cannot collocate, write to a single shard @aly_cabral READS WRITESREADS
- 21. READ CONCERN = READ ISOLATION Options: LOCAL, MAJORITY, SNAPSHOT, LINEARIZABLE Fun Fact: ReadConcern: Majority does not need to coordinate across nodes to return results! @aly_cabral
- 22. Read Concern: Snapshot P S S P S S P S S time: 102 Majority or Local time: 102 time: 110 time: 104 @aly_cabral
- 24. @aly_cabral TXN Write TXN Write {Doc 1} TXN write is unable to obtain a write lock! TXN is ABORTED rolledback entirely
- 25. @aly_cabral TXN Write Operation Non-Transactional Write {Doc 1} Operation Blocks Behind TXN Commit Infinitely retries with Backoff Logic until MaxTimeMS is reached Operation non-transactional write is unable to obtain a write lock!
- 26. @aly_cabral When non-transactional writes encounter write conflicts Operation Blocks Behind TXN Commit Infinitely retries with Backoff Logic until MaxTimeMS is reached When transactions encounter write conflicts Transactions abort rollback entirely
- 27. @aly_cabral TXN Read TXN Write {Doc 1} Reads do not lock a document from being modified Only writes make write conflicts
- 28. More New Functionality in 4.2 @aly_cabral
- 29. Transactions Larger than 16MB In MongoDB 4.0: @aly_cabral {<statement 1>} {<statement 2>} {<statement 3>} {<statement 4>} In MongoDB 4.2: {<statement 1>, <statement 2>, <statement 3>, <statement 4>}
- 30. This does not mean you can have infinitely large transactions -- • WT still has to maintain the history since the snapshot time • Transactions have the same snapshot throughout the duration of the transaction • Transactions automatically abort after 60 seconds – TUNABLE. transactionLifetimeLimitSeconds @aly_cabral
- 31. Mutable Shard Key Values Tiered storage: Aging out older documents to low cost storage shard • Global-redistribution: Rehoming documents to a new region @aly_cabral ShardShard
- 32. • Use MongoDB 4.2 drivers • Transactions targeting a single shard will be faster than those spanning multiple shards • Best practice: No more than 1,000 documents modified in a single transaction • No arbiters • Chunk migrations will block behind running transactions @aly_cabral Parting Wisdom
- 34. • Free ▪ Online ▪ 9 Chapters • Register Now! - https://university.mongodb.com/courses/M042/about @aly_cabral
- 35. Thank You!