flaky test: test meta_node::meta_node_lifecycle::test_meta_node_leave_last_not_allowed #15605
Assignees
Comments
|
Get it. I'll fix it in : |
drmingdrmer
added a commit
to drmingdrmer/databend
that referenced
this issue
May 22, 2024
This commit introduces a significant change in the snapshot management
process by separating the snapshot writing and committing phases:
- **Snapshot Writing Update:**
- The `WriterV002::spawn_writer_thread()` function now solely handles
the writing of snapshot data and flushing it to disk. The
responsibility of moving the snapshot from a temporary location to
the final path has been decoupled from this thread.
- **Enhanced Control Over Snapshot Final Path:**
- This change gives the caller more control over how to construct the
final path of the snapshot. For instance:
- When importing data from a previous version on disk, the final
path is constructed using the `last-applied-log-id` found in the
imported data.
- When building a snapshot from the state machine, the final path is
constructed using the snapshot meta passed in.
- **Introduction of `TempSnapshotData`:**
- A new structure, `TempSnapshotData`, has been introduced to clearly
differentiate between uncommitted temporary snapshot data and
committed visible snapshot data.
Fix flaky test in:
- datafuselabs#15605
drmingdrmer
added a commit
to drmingdrmer/databend
that referenced
this issue
May 22, 2024
This commit introduces a significant change in the snapshot management
process by separating the snapshot writing and committing phases:
- **Snapshot Writing Update:**
- The `WriterV002::spawn_writer_thread()` function now solely handles
the writing of snapshot data and flushing it to disk. The
responsibility of moving the snapshot from a temporary location to
the final path has been decoupled from this thread.
- **Enhanced Control Over Snapshot Final Path:**
- This change gives the caller more control over how to construct the
final path of the snapshot. For instance:
- When importing data from a previous version on disk, the final
path is constructed using the `last-applied-log-id` found in the
imported data.
- When building a snapshot from the state machine, the final path is
constructed using the snapshot meta passed in.
- **Introduction of `TempSnapshotData`:**
- A new structure, `TempSnapshotData`, has been introduced to clearly
differentiate between uncommitted temporary snapshot data and
committed visible snapshot data.
Fix flaky test in:
- datafuselabs#15605
drmingdrmer
added a commit
that referenced
this issue
May 22, 2024
This commit introduces a significant change in the snapshot management
process by separating the snapshot writing and committing phases:
- **Snapshot Writing Update:**
- The `WriterV002::spawn_writer_thread()` function now solely handles
the writing of snapshot data and flushing it to disk. The
responsibility of moving the snapshot from a temporary location to
the final path has been decoupled from this thread.
- **Enhanced Control Over Snapshot Final Path:**
- This change gives the caller more control over how to construct the
final path of the snapshot. For instance:
- When importing data from a previous version on disk, the final
path is constructed using the `last-applied-log-id` found in the
imported data.
- When building a snapshot from the state machine, the final path is
constructed using the snapshot meta passed in.
- **Introduction of `TempSnapshotData`:**
- A new structure, `TempSnapshotData`, has been introduced to clearly
differentiate between uncommitted temporary snapshot data and
committed visible snapshot data.
Fix flaky test in:
- #15605
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Summary
https://github.com/datafuselabs/databend/actions/runs/9184001132/job/25255645607?pr=15602
cc @drmingdrmer
The text was updated successfully, but these errors were encountered: