release-21.1: kvserver: fix write below closedts bug #63861
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
Make it clear that an else case is, in fact, an else case. Release note: None
Clarifiy a <= vs < in a closedts comment. Release note: None
Release note: None
Make this testing filter run after a closed timestamp is assigned to the command to be filtered. I had to slightly change a test because, now, the filter doesn't run if the raft group is nil. Release note: None
The TestServer/TestCluster have facilities for querying a range's lease. This patch addresses two problems with this functionality: 1) If the node being queried has a lease proposal in flight, the respective functions would return it at the detriment of the current lease. This behavior is dubious, and indeed not what a test that I'm writing wants. This patch adds a function that returns both the current and the prospective next lease. 2) The functions let you ask for the state of a particular node to be queried. This was buggy, though, because the LeaseInfoRequest could be transparently forwarded to a different node in case the queried node doesn't have a replica or has a learner replica. That's confusing as hell. This patch optionally dissallows that, giving the caller confidence that they got the state from the node it was interested in. Release note: None
This patch fixes a bug in our closed timestamp management. This bug was making it possible for a command to close a timestamp even though other requests writing at lower timestamps are currently evaluating. The problem was that we were assuming that, if a replica is proposing a new lease, there can't be any requests in flight and every future write evaluated on the range will wait for the new lease and the evaluate above the lease start time. Based on that reasoning, the proposal buffer was recording the lease start time as its assignedClosedTimestamp. This was matching what it does for every write, where assignedClosedTimestamp corresponds to the the closed timestamp carried by the command. It turns out that the replica's reasoning was wrong. It is, in fact, possible for writes to be evaluating on the range when the lease acquisition is proposed. And these evaluations might be done at timestamps below the would-be lease's start time. This happens when the replica has already received a lease through a lease transfer. The transfer must have applied after the previous lease expired and the replica decided to start acquiring a new one. This fixes one of the assertion failures seen in cockroachdb#62655. Release note (bug fix): A bug leading to crashes with the message "writing below closed ts" has been fixed.
|
This change is |
nvanbenschoten
approved these changes
Apr 19, 2021
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.
Backport 6/6 commits from #63672.
/cc @cockroachdb/release
This patch fixes a bug in our closed timestamp management. This bug was
making it possible for a command to close a timestamp even though other
requests writing at lower timestamps are currently evaluating. The
problem was that we were assuming that, if a replica is proposing a new
lease, there can't be any requests in flight and every future write
evaluated on the range will wait for the new lease and the evaluate
above the lease start time. Based on that reasoning, the proposal buffer
was recording the lease start time as its assignedClosedTimestamp. This
was matching what it does for every write, where assignedClosedTimestamp
corresponds to the the closed timestamp carried by the command.
It turns out that the replica's reasoning was wrong. It is, in fact,
possible for writes to be evaluating on the range when the lease
acquisition is proposed. And these evaluations might be done at
timestamps below the would-be lease's start time. This happens when the
replica has already received a lease through a lease transfer. The
transfer must have applied after the previous lease expired and the
replica decided to start acquiring a new one.
This fixes one of the assertion failures seen in #62655.
Release note (bug fix): A bug leading to crashes with the message
"writing below closed ts" has been fixed.