Skip to content
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

Merge dev into v0.6 #742

Merged
merged 6 commits into from
Nov 4, 2019
Merged

Merge dev into v0.6 #742

merged 6 commits into from
Nov 4, 2019

Conversation

smukil
Copy link
Contributor

@smukil smukil commented Nov 4, 2019

No description provided.

Without proper cleanup, we risk leaking memory and having multiple
owners overwriting the same memory.
Setup/cleanup (struct msg)->msg_info correctly
If we have metadata that exists for a key, but the key itself does
not exist, we would previously get an ambiguous response with the
status code 'E' but a (nul) value.

This patchset deletes the metadata if this case is detected and returns
status code 'X' to denote that the value does not exist.

In normal operation, this ambiguous case should never happen, however,
it can happen if read repairs had been disabled and then reenabled
with live traffic between them. In this case, we don't want it to
crash. This was found in a TEST cluster and I've confirmed that this
fixes it.
Make scripts auto-cleanup stale metadata if detected
Turning on repairs with constant traffic runs the risk of in-flight
queries trying to be repaired. This is incorrect and we should
make sure to not take the repair path for these queries.

This was found in a TEST cluster and I confirmed that this patch
fixes it.
Turning on repairs shouldn't invoke repairs for in-flight queries
@smukil smukil self-assigned this Nov 4, 2019
@smukil smukil merged commit e9c94c1 into v0.6 Nov 4, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant