-
Notifications
You must be signed in to change notification settings - Fork 3.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
kv: bypass lease transfer safety checks during joint consensus #89340
kv: bypass lease transfer safety checks during joint consensus #89340
Conversation
Do you think we should add a unit test that asserts that the lease transfer to a dead VOTER_INCOMING works? |
c387857
to
ec37e5e
Compare
Yes, an end-to-end unit test was needed here. We were already testing that the new flag was properly set in a few tests, but not that this allowed the range to exit a joint configuration when the VOTER_INCOMING is dead. I've added a test in a new commit. It's flaky now without #89221 because the VOTER_DEMOTING is allowed to re-acquire the lease (the starvation we hypothesized), which @shralex is working on. We'll need to sequence these changes appropriately. |
TFTR! I'll have to wait for #89564 to merge before this will pass CI. |
This PR restricts the case when a VOTER_DEMOTING_LEARNER can aquire the lease in a joint configuration to only the case where it was the last leaseholder. Since it is being removed, we only want it to get the lease if no other replica can aquire it, the scenario described in cockroachdb#83687 This fix solves a potential starvation scenario where a VOTER_DEMOTING_LEARNER keeps transferring the lease to the VOTER_INCOMING, succeeding, but then re-acquiring because the VOTER_INCOMING is dead and the lease expires. In this case, we would want another replica to pick up the lease, which would allow us to exit the joint configuration. This PR also removes the leaseHolderRemovalAllowed parameter of CheckCanReceiveLease, since it is always true since 22.2. Release note (bug fix): narrows down the conditions under which a VOTER_DEMOTING_LEARNER can acquire the lease in a joint configuration to a) there has to be an VOTER_INCOMING in the configuration and b) the VOTER_DEMOTING_LEARNER was the last leaseholder. This prevents it from acquiring the lease unless it is the only one that can acquire it, because transferring the lease away is necessary before exiting the joint config (without the fix the system can be stuck in a joint configuration in some rare situations). Fixes: cockroachdb#88667 See also cockroachdb#89340
This PR restricts the case when a VOTER_DEMOTING_LEARNER can aquire the lease in a joint configuration to only the case where it was the last leaseholder. Since it is being removed, we only want it to get the lease if no other replica can aquire it, the scenario described in cockroachdb#83687 This fix solves a potential starvation scenario where a VOTER_DEMOTING_LEARNER keeps transferring the lease to the VOTER_INCOMING, succeeding, but then re-acquiring because the VOTER_INCOMING is dead and the lease expires. In this case, we would want another replica to pick up the lease, which would allow us to exit the joint configuration. This PR also removes the leaseHolderRemovalAllowed parameter of CheckCanReceiveLease, since it is always true since 22.2. Release note (bug fix): narrows down the conditions under which a VOTER_DEMOTING_LEARNER can acquire the lease in a joint configuration to a) there has to be an VOTER_INCOMING in the configuration and b) the VOTER_DEMOTING_LEARNER was the last leaseholder. This prevents it from acquiring the lease unless it is the only one that can acquire it, because transferring the lease away is necessary before exiting the joint config (without the fix the system can be stuck in a joint configuration in some rare situations). Fixes: cockroachdb#88667 See also cockroachdb#89340
89564: kvserver: don't allow VOTER_DEMOTING to acquire lease after transfer r=shralex a=shralex This PR restricts the case when a VOTER_DEMOTING_LEARNER can aquire the lease in a joint configuration to only the case where it was the last leaseholder. Since it is being removed, we only want it to get the lease if no other replica can aquire it, the scenario described in #83687 This fix solves a potential starvation scenario where a VOTER_DEMOTING_LEARNER keeps transferring the lease to the VOTER_INCOMING, succeeding, but then re-acquiring because the VOTER_INCOMING is dead and the lease expires. In this case, we would want another replica to pick up the lease, which would allow us to exit the joint configuration. This PR also removes the leaseHolderRemovalAllowed parameter of CheckCanReceiveLease, since it is always true since 22.2. Release note (bug fix): narrows down the conditions under which a VOTER_DEMOTING_LEARNER can acquire the lease in a joint configuration to a) there has to be an VOTER_INCOMING in the configuration and b) the VOTER_DEMOTING_LEARNER was the last leaseholder. This prevents it from acquiring the lease unless it is the only one that can acquire it, because transferring the lease away is necessary before exiting the joint config (without the fix the system can be stuck in a joint configuration in some rare situations). Fixes: #88667 See also #89340 Co-authored-by: shralex <[email protected]>
This PR restricts the case when a VOTER_DEMOTING_LEARNER can aquire the lease in a joint configuration to only the case where it was the last leaseholder. Since it is being removed, we only want it to get the lease if no other replica can aquire it, the scenario described in #83687 This fix solves a potential starvation scenario where a VOTER_DEMOTING_LEARNER keeps transferring the lease to the VOTER_INCOMING, succeeding, but then re-acquiring because the VOTER_INCOMING is dead and the lease expires. In this case, we would want another replica to pick up the lease, which would allow us to exit the joint configuration. This PR also removes the leaseHolderRemovalAllowed parameter of CheckCanReceiveLease, since it is always true since 22.2. Release note (bug fix): narrows down the conditions under which a VOTER_DEMOTING_LEARNER can acquire the lease in a joint configuration to a) there has to be an VOTER_INCOMING in the configuration and b) the VOTER_DEMOTING_LEARNER was the last leaseholder. This prevents it from acquiring the lease unless it is the only one that can acquire it, because transferring the lease away is necessary before exiting the joint config (without the fix the system can be stuck in a joint configuration in some rare situations). Fixes: #88667 See also #89340
This PR restricts the case when a VOTER_DEMOTING_LEARNER can aquire the lease in a joint configuration to only the case where it was the last leaseholder. Since it is being removed, we only want it to get the lease if no other replica can aquire it, the scenario described in #83687 This fix solves a potential starvation scenario where a VOTER_DEMOTING_LEARNER keeps transferring the lease to the VOTER_INCOMING, succeeding, but then re-acquiring because the VOTER_INCOMING is dead and the lease expires. In this case, we would want another replica to pick up the lease, which would allow us to exit the joint configuration. This PR also removes the leaseHolderRemovalAllowed parameter of CheckCanReceiveLease, since it is always true since 22.2. Release note (bug fix): narrows down the conditions under which a VOTER_DEMOTING_LEARNER can acquire the lease in a joint configuration to a) there has to be an VOTER_INCOMING in the configuration and b) the VOTER_DEMOTING_LEARNER was the last leaseholder. This prevents it from acquiring the lease unless it is the only one that can acquire it, because transferring the lease away is necessary before exiting the joint config (without the fix the system can be stuck in a joint configuration in some rare situations). Fixes: #88667 See also #89340
This PR restricts the case when a VOTER_DEMOTING_LEARNER can aquire the lease in a joint configuration to only the case where it was the last leaseholder. Since it is being removed, we only want it to get the lease if no other replica can aquire it, the scenario described in cockroachdb#83687 This fix solves a potential starvation scenario where a VOTER_DEMOTING_LEARNER keeps transferring the lease to the VOTER_INCOMING, succeeding, but then re-acquiring because the VOTER_INCOMING is dead and the lease expires. In this case, we would want another replica to pick up the lease, which would allow us to exit the joint configuration. Release note (bug fix): narrows down the conditions under which a VOTER_DEMOTING_LEARNER can acquire the lease in a joint configuration to a) there has to be an VOTER_INCOMING in the configuration and b) the VOTER_DEMOTING_LEARNER was the last leaseholder. This prevents it from acquiring the lease unless it is the only one that can acquire it, because transferring the lease away is necessary before exiting the joint config (without the fix the system can be stuck in a joint configuration in some rare situations). Release justification: solves a serious bug. Fixes: cockroachdb#88667 See also cockroachdb#89340
Also simplify a bit of logic.
Fixes cockroachdb#88667. This commit adds logic to bypass lease transfer safety checks (added in 034611b) when in a joint configuration and transferring the lease from a VOTER_DEMOTING to a VOTER_INCOMING. We do so because we could get stuck without a path to exit the joint configuration if we rejected this lease transfer while waiting to confirm that the target is up-to-date on its log. That confirmation may never arrive if the target is dead or partitioned away, and while we'd rather not transfer the lease to a dead node, at least we have a mechanism to recovery from that state. We also just sent the VOTER_INCOMING a snapshot (as a LEARNER, before promotion), so it is unlikely that the replica is actually dead or behind on its log. A better alternative here would be to introduce a mechanism to choose an alternate lease transfer target after some amount of time, if the lease transfer to the VOTER_INCOMING cannot be confirmed to be safe. We may do this in the future, but given the proximity to the release and given that this matches the behavior in v22.1, we choose this approach for now. Release note: None Release justification: Needed to resolve release blocker.
This commit adds a new test which ensures that the lease transfer performed during a joint config replication change that is replacing the existing leaseholder does not get stuck even if the existing leaseholder cannot prove that the incoming leaseholder is caught up on its log. It does so by killing the incoming leaseholder before it receives the lease and ensuring that the range is able to exit the joint configuration. Currently, the range exits by bypassing safety checks during the lease transfer, sending the lease to the dead incoming voter, letting the lease expire, acquiring the lease on one of the non-demoting voters, and exiting. The details here may change in the future, but the goal of this test will not.
0fd4251
to
f29f178
Compare
bors r=shralex |
Build succeeded: |
Encountered an error creating backports. Some common things that can go wrong:
You might need to create your backport manually using the backport tool. error creating merge commit from 0036160 to blathers/backport-release-22.2-89340: POST https://api.github.com/repos/cockroachdb/cockroach/merges: 409 Merge conflict [] you may need to manually resolve merge conflicts with the backport tool. Backport to branch 22.2.x failed. See errors above. error creating merge commit from 0036160 to blathers/backport-release-22.2.0-89340: POST https://api.github.com/repos/cockroachdb/cockroach/merges: 409 Merge conflict [] you may need to manually resolve merge conflicts with the backport tool. Backport to branch 22.2.0 failed. See errors above. 🦉 Hoot! I am a Blathers, a bot for CockroachDB. My owner is otan. |
Pardon if this was already discussed, should this backport to v22.1? |
These lease transfer safety checks were only added in v22.2 (in 034611b), so this doesn't need a backport. We were hopeful in #82758 that the change would be non-disruptive and eventually qualify for a backport, but as with many changes in this area, the effects are far-reaching, so we did not gain enough confidence to meet the backport criteria. |
This commit adds logic to bypass lease transfer safety checks (added in 034611b) when in a joint configuration and transferring the lease from a VOTER_DEMOTING to a VOTER_INCOMING. We do so because we could get stuck without a path to exit the joint configuration if we rejected this lease transfer while waiting to confirm that the target is up-to-date on its log. That confirmation may never arrive if the target is dead or partitioned away, and while we'd rather not transfer the lease to a dead node, at least we have a mechanism to recovery from that state. We also just sent the VOTER_INCOMING a snapshot (as a LEARNER, before promotion), so it is unlikely that the replica is actually dead or behind on its log.
A better alternative here would be to introduce a mechanism to choose an alternate lease transfer target after some amount of time, if the lease transfer to the VOTER_INCOMING cannot be confirmed to be safe. We may do this in the future, but given the proximity to the release and given that this matches the behavior in v22.1, we choose this approach for now.
Release note: None
Release justification: Needed to resolve release blocker.
Fixes: #88667
See also #89564