You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Feb 16, 2022. It is now read-only.
When using the logspec client, the cluster needs to regularly calculate a safe truncation point, and tell the log store to truncate the prefix of the log to that index.
The API server can consult with all the views to determine the smallest safe point across the cluster. Either the API tier can then issue the delete, or for more complex deployments where a single log is used across multiple DCs, and each DC has its own beam cluster, you'd probably want this truncation to be managed by a separate control plane.
The text was updated successfully, but these errors were encountered:
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
When using the logspec client, the cluster needs to regularly calculate a safe truncation point, and tell the log store to truncate the prefix of the log to that index.
The API server can consult with all the views to determine the smallest safe point across the cluster. Either the API tier can then issue the delete, or for more complex deployments where a single log is used across multiple DCs, and each DC has its own beam cluster, you'd probably want this truncation to be managed by a separate control plane.
The text was updated successfully, but these errors were encountered: