Matt Butrovich

MVTO: Allow reads of write locked tuples as long as reader's TS < writer's TS

1

Currently we abort if we try to read a tuple whose write lock is held by another transaction. This is too conservative. We can read the tuple if the txn...

Performance fix: replace WorkerPool sleeping with condition variable

3

I couldn't stare at concurrency control logic anymore this afternoon, so I threw together a quick experiment replacing our WorkerPool's sleep with exponential backoff behavior for [C++11's std::condition_variable](https://en.cppreference.com/w/cpp/thread/condition_variable). This change...

We've found a number of tests (serializable_transaction_test, for example) that sleep a thread for exactly the duration of an epoch and expect the global epoch to have advanced. However, this...

Concurrency control (MVTO) discussion

1

In trying to improve the performance of the CC system, I've identified several design decisions that I'd like to discuss here and possibly change in the future. 1. We maintain...

David and I completed our first comprehensive evaluation of the garbage collector. Based on the test suite we wrote [here](https://github.com/mbutrovich/peloton/blob/gc_fixes/test/gc/transaction_level_gc_manager_test.cpp), we've identified the following tuple-related problem scenarios: 1. Primary key...

End-to-end (JUnit and oltpbench) only for coverage

1

This is a test PR just to see what coverage looks like if we only run JUnit and oltpbench.

Spotted on #1617 which makes an unrelated change. ``` [ RUN ] ChainingHashTableTest.SimpleIteration /jenkins/workspace/terrier_PR-1617/test/execution/sql_chaining_hash_table_test.cpp:284: Failure Expected equality of these values: reference.size() Which is: 499 found_entries Which is: 500 /jenkins/workspace/terrier_PR-1617/test/execution/sql_chaining_hash_table_test.cpp:305: Failure...

Unhack DROP ... IF EXISTS support by adding code to the binder

1

`DROP` statements allow for an `IF EXISTS ` with the following semantics: > Do not throw an error if the [object] does not exist. A notice is issued in this...

Running [CH-benCHmark](https://db.in.tum.de/research/projects/CHbenCHmark/index.shtml?lang=en) via [oltpbench](https://github.com/oltpbenchmark/oltpbench/tree/master/src/com/oltpbenchmark/benchmarks/chbenchmark). Using the attached schemas for TPC-C and chbenchmark tables, and attached config. [chbenchmark.zip](https://github.com/cmu-db/noisepage/files/6511230/chbenchmark.zip) ``` ant execute -Dconfig=./config/noisepage_tpcc_config.xml \ -Dbenchmark=tpcc -Dexecute=false -Dload=true -Dcreate=true ant execute -Dconfig=./config/noisepage_ch_config.xml \...

script.testing.replication.tests_simple failing

2

Seen it on PRs and on master a few times since #1472 merged. Example from most recent master build: http://jenkins.db.cs.cmu.edu:8080/blue/organizations/jenkins/terrier/detail/master/921/pipeline I know @lmwnshn is aware and watching it, but just...