Don't pass InvalidTransactionId to update_next_xid. #6410
Merged
Conversation
This file contains 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
kelvich
approved these changes
Jan 20, 2024
2268 tests run: 2179 passed, 0 failed, 89 skipped (full report)Code coverage (full report)
The comment gets automatically updated with the latest test results
4a8d33e at 2024-01-20T14:47:46.807Z :recycle: |
|
hlinnaka
force-pushed
the
fix-update_next_xid
branch
from
6ee90cf
to
5da65ec
Compare
kelvich
force-pushed
the
fix-update_next_xid
branch
from
5da65ec
to
771f9d3
Compare
update_next_xid() doesn't have any special treatment for the invalid or other special XIDs, so it will treat InvalidTransactionId (0) as a regular XID. If old nextXid is smaller than 2^31, 0 will look like a very old XID, and nothing happens. But if nextXid is greater than 2^31 0 will look like a very new XID, and update_next_xid() will incorrectly bump up nextXID.
kelvich
force-pushed
the
fix-update_next_xid
branch
from
771f9d3
to
f10adb6
Compare
hlinnaka
force-pushed
the
fix-update_next_xid
branch
from
e19df29
to
6a2187f
Compare
Without this PR, when importing a vanilla Postgres datadir with nextXid < 1024, importing the CHECKPOINT record would bump it up to 1024. That caused a new delta layer to be created at that point. After fixing that, if the WAL that's part of the basebackup didn't contain any updates to actual relations, importing the WAL would generate no delta layers. That caused a test failure, because the test waited for the generated delta layers to be uploaded, but none were generated. To fix, update the checkpoint key-value pair on every CHECKPOINT record. That's not really required, but seems nice anyway, and it forces a delta layer to be generated in this corner case.
hlinnaka
commented
Jan 20, 2024
|
One test failure in the e2e public API tests: https://github.com/neondatabase/cloud/actions/runs/7595014557/job/20687120459. I'm pretty sure that's a flaky test, because all the other variants of the e2e tests failed, and they don't exercise this codepath in any special manner anyway. I will relaunch it. |
|
The e2e test suite failed again, at a different test. I still believe it's flakiness, but would appreciate a second opinion from someone more in touch with the e2e tests |
5 tasks
abhigets
pushed a commit
that referenced
this pull request
Jan 24, 2024
update_next_xid() doesn't have any special treatment for the invalid or other special XIDs, so it will treat InvalidTransactionId (0) as a regular XID. If old nextXid is smaller than 2^31, 0 will look like a very old XID, and nothing happens. But if nextXid is greater than 2^31 0 will look like a very new XID, and update_next_xid() will incorrectly bump up nextXID.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
update_next_xid() doesn't have any special treatment for the invalid or other special XIDs, so it will treat InvalidTransactionId (0) as a regular XID. If old nextXid is smaller than 2^31, 0 will look like a very old XID, and nothing happens. But if nextXid is greater than 2^31 0 will look like a very new XID, and update_next_xid() will incorrectly bump up nextXID.