fix: read/write splitting switch to writer #1080
Open
+160
−42
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
faviansamatha
force-pushed
the
fix/read-write-splitting-no-writer
branch
from
9217b12 to
f3c0000
Compare
sophia-bq
reviewed
Jan 20, 2026
| props["custom_endpoint_idle_monitor_expiration_ms"] = 30_000 | ||
| props["wait_for_custom_endpoint_info_timeout_ms"] = 30_000 | ||
|
|
||
| # Ensure that we are starting with a reader connection |
Contributor
There was a problem hiding this comment.
Suggested change
| # Ensure that we are starting with a reader connection | |
| # Ensure that we are starting with a writer connection |
| try: | ||
| self.wait_until_endpoint_has_members(rds_client, {original_instance_id, new_member}) | ||
|
|
||
| # We should now be able to switch to reader. |
Contributor
There was a problem hiding this comment.
Suggested change
| # We should now be able to switch to reader. | |
| # We should now be able to switch to writer. |
| conn.read_only = False | ||
|
|
||
| writer_id = rds_utils.get_cluster_writer_instance_id() | ||
| new_member = writer_id |
Contributor
There was a problem hiding this comment.
why do we assign writer_id to new_member? the following lines are actually clearer if you keep the name writer_id?
Comment on lines
+284
to
+285
| # We should not be able to switch again because new_member was removed from the custom endpoint. | ||
| # We are connected to the reader. Attempting to switch to the writer will throw an exception. |
Contributor
There was a problem hiding this comment.
To clarify:
- reconnected back to the original reader instance on line 277
- removed the original reader instance from the custom endpoint
- now switching to the writer instance (a valid member of the endpoint) errors out
This is because the current reader connection is invalid? Wouldn't any operation fail?
| assert new_instance_id == new_member | ||
|
|
||
| # Switch back to original instance | ||
| conn.read_only = True |
Contributor
There was a problem hiding this comment.
Suggested change
| conn.read_only = True | |
| conn.read_only = True | |
| new_instance_id = rds_utils.query_instance_id(conn) | |
| assert new_instance_id == original_instance_id |
| with pytest.raises(ReadWriteSplittingError): | ||
| conn.read_only = new_read_only_value | ||
| # We are connected to the writer. Attempting to switch to the reader will not work but will intentionally | ||
| # not throw an exception. In this scenario we log a warning and purposefully stick with the writer. |
Contributor
There was a problem hiding this comment.
Suggested change
| # not throw an exception. In this scenario we log a warning and purposefully stick with the writer. | |
| # not throw an exception. In this scenario we log a warning and fallback to the writer. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.