-
Type:
Bug
-
Resolution: Duplicate
-
Priority:
Major - P3
-
None
-
Affects Version/s: None
-
Component/s: None
-
Replication
-
Execution Team 2025-02-03
Attempting to execute a transaction which updates a number of documents that causes the WiredTiger dirty cache ratio to significantly exceed 20% results in a write conflict error instead of a transaction too large for cache error.
In testing on an Replica Set, I was unable to trigger a transactionTooLargeForCache error even though FTDC reported that the cluster's dirty cache ratio exceeded 90% and there was no other additional workload happening.
Repro
Dependencies:
- Python 3
- mongo-mangler (https://github.com/deadsix/mongo-mangler/tree/add-more-pipelines)
- pymongo
// Create Collection via mongo-mangler from Command Line python3 mongo-mangler.py -m mongodb+srv://myDatabaseUser:D1fficultP%40ssw0rd@cluster0.example.mongodb.net/ -s 500000 -p ~/mongo_mangler/mongo-mangler-add-more-pipelines/examples/pipeline_example_fake_accounts_one_kb_doc.js //Connect to Cluster via mongosh [primary] test> session = db.getMongo().startSession( { readPreference: { mode: "primary" } } ); [primary] test> session.startTransaction( { readConcern: { level: "local" }, writeConcern: { w: "majority" } } ); coll = session.getDatabase("test").big_collection;try { coll.updateMany( {dependent_count: 1}, {$set: {recent_payment: 2421.24}} ); } catch (error) { session.abortTransaction(); print(error); }
- duplicates
-
-
- Open
-
- related to
-
-
- Open
-
-
SERVER-94010 Delete transaction_too_large_for_cache/temporarily_unavailable_on_secondary_transaction_application.js and replace with unit test
-
- Closed
-