You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
the above sql will result in a channel close error, which users can not understand
the root casue found in CN, the pk is null and make sink exec panic
{"timestamp":"2023-10-18T03:03:34.181682374Z","level":"ERROR","fields":{"message":"failed to collect barrier: Actor 29 exit unexpectedly: Executor error: Sink error: Remote sink error: gRPC error (Internal error): Error when exec INSERT INTO public.room_user_visit_1min(report_ts, uv, room_id) VALUES (NULL, 1, '1364688900') ON CONFLICT (report_ts, room_id) DO UPDATE SET report_ts=EXCLUDED.report_ts, uv=EXCLUDED.uv, room_id=EXCLUDED.room_id
RETURNING *, message ERROR: null value in column \"report_ts\" of relation \"room_user_visit_1min\" violates not-null constraint
Detail: Failing row contains (null, 1364688900, 1).
backtrace of `StreamExecutorError`:
0: <risingwave_stream::executor::error::Inner as core::convert::From<risingwave_stream::executor::error::ErrorKind>>::from
at ./risingwave/src/stream/src/executor/error.rs:40:10
1: <T as core::convert::Into<U>>::into
at ./rustc/f0411ffcebcd7f75ac02ed45feb53ffd07b75398/library/core/src/convert/mod.rs:717:9
2: <risingwave_stream::executor::error::StreamExecutorError as core::convert::From<risingwave_stream::executor::error::ErrorKind>>::from
at ./risingwave/src/stream/src/executor/error.rs:143:34
3: <T as core::convert::Into<U>>::into
at ./rustc/f0411ffcebcd7f75ac02ed45feb53ffd07b75398/library/core/src/convert/mod.rs:717:9
4: <risingwave_stream::executor::error::StreamExecutorError as core::convert::From<risingwave_connector::sink::SinkError>>::from
at ./risingwave/src/stream/src/executor/error.rs:209:33
5: <core::result::Result<T,F> as core::ops::try_trait::FromResidual<core::result::Result<core::convert::Infallible,E>>>::from_residual
at ./rustc/f0411ffcebcd7f75ac02ed45feb53ffd07b75398/library/core/src/result.rs:1961:27
6: risingwave_stream::executor::sink::SinkExecutor<F>::execute_consume_log::{{closure}}
at ./risingwave/src/stream/src/executor/sink.rs:279:25
Error message/log
No response
To Reproduce
No response
Expected behavior
No response
How did you deploy RisingWave?
No response
The version of RisingWave
No response
Additional context
No response
The text was updated successfully, but these errors were encountered:
Kernel version 1.2.0.
Sink error causes the Actor exit, then trigger the recovery procedure which will reset the compute node. This is by design w/o sink decouple.
had a discussion with @st1page, the expected behavior is to discard the record with NULL pk, and report a warning. This rule applies to all connector x format with the required primary_key.
The only problem left is kinesis x append-only, kinesis requires a message key but it does not mean anything in append-only, what do you think @xiangjinwu
You meant as a natural extension kinesis x append-only would also discard records with nulls in pk, but it is also okay (maybe preferred) not to discard them?
As long as it writes to a message queue, having nulls in some pk columns (including all nulls) is not as severe as inserts into a DB. The actual meaning is partition key #9768.
After some discussion, I think it is a expected behavior. We've proposed log store to handle such case. And we don't need to add extra behavior than let it stuck on.
Describe the bug
the above sql will result in a channel close error, which users can not understand
the root casue found in CN, the pk is null and make sink exec panic
Error message/log
No response
To Reproduce
No response
Expected behavior
No response
How did you deploy RisingWave?
No response
The version of RisingWave
No response
Additional context
No response
The text was updated successfully, but these errors were encountered: