Shankar

Results 25 comments of Shankar

More details about problem cause in respective tickets : https://github.com/ClickHouse/ClickHouse/issues/68990 and https://github.com/ClickHouse/ClickHouse/issues/68988

Please move PR to testing if all good, thanks!

Checked the failure : ``` test_role/test.py::test_roles_cache[1-5] [gw0] [ 86%] PASSED test_role/test.py::test_roles_cache[1-5] test_role/test.py::test_roles_cache[2-5] [gw0] [ 87%] PASSED test_role/test.py::test_roles_cache[2-5] test_role/test.py::test_roles_cache[3-5] [gw0] [ 89%] PASSED test_role/test.py::test_roles_cache[3-5] test_role/test.py::test_roles_cache[4-5] [gw0] [ 90%] PASSED test_role/test.py::test_roles_cache[4-5] test_role/test.py::test_roles_cache[5-5]...

Test run is successful after adding a `node1.start_clickhouse` at end of testcase.

>the problem is the extra requests, not the extra nodes Yes, the extra requests are the problem, explained here : https://github.com/ClickHouse/ClickHouse/issues/68990 I have added `use_cluster=false` to the XML files as...

Even with files cleanup, I think the 2nd iteration of the test (without a full teardown/restart) will save and upload snapshots named : `snapshot_250.bin.zstd, snapshot_300.bin.zstd...`?

Right, I have run it locally a few times to understand and I saw that the snapshots on the local disk are automatically being purged and the numbering keeps moving...

I did the code changes and tested many times, but I am unable to get determinism for iteration 2 and beyond. ``` 2024-09-05 04:55:45 [ 417 ] WARNING : assert...