Download as PDF, PPTX
![lcov [ 1 / 2]](https://image.slidesharecdn.com/quality-assurance-in-postgresql-v2-170428100547/75/Quality-Assurance-in-PostgreSQL-21-2048.jpg)
![lcov [ 2 / 2 ]](https://image.slidesharecdn.com/quality-assurance-in-postgresql-v2-170428100547/75/Quality-Assurance-in-PostgreSQL-22-2048.jpg)
Uploaded byAleksander Alekseev
Quality Assurance in PostgreSQL
This document discusses quality assurance processes for PostgreSQL, an open source object-relational database. It covers development practices like code reviews on mailing lists and testing patches through "commitfests". Comprehensive testing is emphasized, including modular unit tests, integration tests, performance benchmarks, and documentation reviews. A variety of tools are also presented to aid testing, debugging, and performance profiling of PostgreSQL code.
Quality Assurance in PostgreSQL
- 1. Quality Assurance in PostgreSQL AnastasiaLubennikova Aleksander Alekseev
- 2. Agenda ● Development ● Testing ●Benchmarking ● Tools ● Other Topics
- 3. What is PostgreSQL? ●Open source object-relational database system ● Over 20 years of development ● Hundreds of developers around the world ● Several major vendors
- 5. What’s so SpecialAbout PostgreSQL? ● Pure C ● Complicated project ● Cross-platform (Windows, Linux, MacOS, Solaris, *BSD, AIX, HP-UX, ...) ● High code quality ● Great documentation ● Major releases every year
- 6. Development Process ● Git,C, Autotools, Perl ● Mailing Lists ○ pgsql-hackers@ ○ pgsql-committers@ ○ pgsql-bugs@ ○ etc ● Conferences ○ pgcon.org ○ pgconf.us ○ pgconf.ru ○ … and many more!
- 7.
- 11. Submitting a Patch ●Optional: Proposal and discussion ● Optional: PoC ● Patch is added to commitfest ● While(not ready): ○ Code review ○ Testing ○ Discussion ○ More discussion ○ Rewrite a patch ● Either: ○ Patch is accepted ○ Patch is moved to the next commitfest ○ Patch is returned with a feedback
- 13. What to test? ●Modular* tests ○ Indexes ○ Replication ○ Types ○ WAL / Recovery ● Integration/system tests ○ How various features work together? ○ Upgrade, Backup/Restore, etc ● Performance ○ Previous versions ○ Other implementations ● Documentation (SGML) ● Alpha and beta tests
- 14. How to test? ●Code review ● Regular SQL input / output tests ○ Can’t test utilities, concurrency, etc.. ○ Must be deterministic ● TAP/Perl tests ○ SSL ○ Replication and recovery ● pgTAP framework ● Testgres - python framework
- 15. SQL tests ● Canonly test SQL-level stuff ● Must be deterministic ● Should be executed to see the exact output
- 16. Benchmarks ● Industry standardbenchmarks ○ TPC-H ○ TPC-E ○ TPC-B ● Pgbench
- 17. Basic Tools ● gcc/ clang / ... ● gdb / lldb ● vim + :set spell spelllang=en_us,ru_ru ● Asserts ● wal_consistency_checking = 'all' ● log_level = DEBUG ● amcheck (property based tests) ● sqlsmith (fuzzing) ● valgrind
- 18.
- 19.
- 20. HeapTrack + MassifVisualizer
- 21. lcov [ 1/ 2]
- 22. lcov [ 2/ 2 ]
- 23.
- 24.
- 25. Links ● https://www.postgresql.org/ ● http://postgres-edu.blogspot.ru/search/label/Hacking%20PostgreSQL ●https://github.com/afiskon/pgscripts ● https://habr.ru/p/308442/ + related articles
- 26. Thank you foryour attention! ● https://postgrespro.com/ (We are hiring! ;) ● a.lubennikova@postgrespro.ru ● a.alekseev@postgrespro.ru