Sema Checherinda
Sema Checherinda
Shot read: lets postpone logging tls_sni until we really need this. Long read: tls_sni (ssl servername) is send by client to the server at the first hello-message during ssl handshake....
> And what's implemented in #38581? :) There is nothing related to the tls_sni, unfortunately. Peer address has been written as a column tls_sni in query_log table. > Patching Poco...
This happens when select from system.parts is called. This select reads all parts, deleted parts also are read. Here https://github.com/ClickHouse/ClickHouse/pull/41145 I encountered this race and made fix for it. Just...
The test [01825_type_json_schema_race_long](https://s3.amazonaws.com/clickhouse-test-reports/41145/c33c4ec4f8b26376c42f845e42160e6c0a526cf8/stateless_tests__aarch64_.html) is failed. I investigate it. At first look it seems like Object(‘json’) has to be wrapped in Tuple. This PR seemed to be related #41290 UPD: test...
Failed 01167_isolation_hermitage https://s3.amazonaws.com/clickhouse-test-reports/41145/ba00c29338d84993603f81a045e75a9021d3eaea/stateless_tests__ubsan_.html The old parts has been held until drop table. Accidentally happen that debug logging wasn't print anything du to pressure timeout in 60 sec. Will make it...
https://github.com/ClickHouse/ClickHouse/issues/41500#issuecomment-1282285537 asan failed -- https://github.com/ClickHouse/ClickHouse/issues/41500
01130_in_memory_parts_partitons part_in_memory is flushed to disk to the temp directory with prefix "tmp_freeze_". That tmp directory just left. The same problem was with "tmp_clone_" before.
stress tests under msan https://s3.amazonaws.com/clickhouse-test-reports/41145/ba00c29338d84993603f81a045e75a9021d3eaea/stress_test__msan_.html the similar stacks as in https://github.com/ClickHouse/ClickHouse/issues/41831
https://s3.amazonaws.com/clickhouse-test-reports/57205/82270627ed1c74ce96c1eb8100430de3b81b3efe/fuzzer_astfuzzermsan/report.html
Don't you mind if I ask some questions for better mine understanding and further discussions? Does the keeper client store the availability flag/information for each server host? That information is...