mirror of https://github.com/facebook/rocksdb.git
15d9f28da5
Summary: Add crash test for the case of best-efforts recovery. After a certain amount of time, we kill the db_stress process, randomly delete some certain table files and restart db_stress. Given the randomness of file deletion, it is difficult to verify against a reference for data correctness. Therefore, we just check that the db can restart successfully. Pull Request resolved: https://github.com/facebook/rocksdb/pull/6819 Test Plan: ``` ./db_stress -best_efforts_recovery=true -disable_wal=1 -reopen=0 ./db_stress -best_efforts_recovery=true -disable_wal=0 -skip_verifydb=1 -verify_db_one_in=0 -continuous_verification_interval=0 make crash_test_with_best_efforts_recovery ``` Reviewed By: anand1976 Differential Revision: D21436753 Pulled By: riversand963 fbshipit-source-id: 0b3605c922a16c37ed17d5ab6682ca4240e47926 |
||
|---|---|---|
| .. | ||
| CMakeLists.txt | ||
| batched_ops_stress.cc | ||
| cf_consistency_stress.cc | ||
| db_stress.cc | ||
| db_stress_common.cc | ||
| db_stress_common.h | ||
| db_stress_driver.cc | ||
| db_stress_driver.h | ||
| db_stress_env_wrapper.h | ||
| db_stress_gflags.cc | ||
| db_stress_listener.h | ||
| db_stress_shared_state.cc | ||
| db_stress_shared_state.h | ||
| db_stress_stat.h | ||
| db_stress_test_base.cc | ||
| db_stress_test_base.h | ||
| db_stress_tool.cc | ||
| no_batched_ops_stress.cc | ||