cauchy1988 commented on issue #2283: URL: https://github.com/apache/incubator-pegasus/issues/2283#issuecomment-4700708070
This looks like the same root cause as #2215: when a replica reopens an incomplete/corrupted RocksDB `rdb` directory, it hits the hard `CHECK_PREFIX_MSG(!missing_meta_cf, ...)` abort, dropping the replica server (`get_alive_replica_server_count()` returns 2) and failing the whole `base_api_test`. PR #2404 replaces the hard abort with graceful degradation — on detecting a missing column family it removes the corrupted `rdb` directory, recreates a fresh DB, and re-syncs from the primary. With that change applied, both corrupt-DB test cases now pass: ``` [ RUN ] integration_test.write_corrupt_db [ OK ] integration_test.write_corrupt_db (40560 ms) [ RUN ] integration_test.read_corrupt_db [ OK ] integration_test.read_corrupt_db (2706 ms) ``` So #2404 should close this issue as well. A review would be appreciated. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
