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
Now, when not running in cluster-aware mode, the directory name for the topic will be prepended with '.', and only when all partitions for that topic have been initialized (in that .directory), that .directory will be atomically renamed to the correct name(i.e the topic name).
This is important because it's possible that partition creation may fail(e.g system fault, FS fault, etc), and we don't want to leave a mess when that happens. That is to say, creating a topic should be an atomic operation.
implement cleanup() lambda in Service::process_create_topic()
implement the same idea in Service::reconcile_cluster_topology()
The text was updated successfully, but these errors were encountered:
Turns out that deleting directories on startup may lead to data loss - because TANK may be misconfigured to use the wrong base directory. Will now generate a warning and will eventually accept a new command-line argument enabling GC on startup.
Now, when not running in cluster-aware mode, the directory name for the topic will be prepended with '.', and only when all partitions for that topic have been initialized (in that .directory), that .directory will be atomically renamed to the correct name(i.e the topic name).
This is important because it's possible that partition creation may fail(e.g system fault, FS fault, etc), and we don't want to leave a mess when that happens. That is to say, creating a topic should be an atomic operation.
Service::process_create_topic()
Service::reconcile_cluster_topology()
The text was updated successfully, but these errors were encountered: