Cache pool config data format Currently this data format is only used by ZookeeperCachePoolManager to read the config data from zookeeper serverset parent node, and the expected cache pool size is the only attribute we need for now.
Cache pool based on a static list
A com.twitter.finagle.Resolver for resolving destination names associated with Twitter cache pools.
Indicates that an error occurred while resolving a cache address.
Indicates that an error occurred while resolving a cache address. See com.twitter.finagle.memcached.TwitterCacheResolver for details.
Zookeeper based cache node group with a serverset as the underlying pool.
Zookeeper based cache node group with a serverset as the underlying pool. It will monitor the underlying serverset changes and report the detected underlying pool size. It will monitor the serverset parent node for cache pool config data, cache node group update will be triggered whenever cache config data change event happens.
Zookeeper based cache pool cluster with a serverset as the underlying pool.
Zookeeper based cache pool cluster with a serverset as the underlying pool. It will monitor the underlying serverset changes and report the detected underlying pool size. It will also monitor the serverset parent node for cache pool config data, cache pool cluster update will be triggered whenever cache config data change event happens.
A zk monitor trait that assists with monitoring a given zk path for any node data change, in which the provided zk data handling implementation will be invoked.
A zk monitor trait that assists with monitoring a given zk path for any node data change, in which the provided zk data handling implementation will be invoked.
This monitor will maintain a queue so that every work item triggered by zk event will be processed in an order with a back off policy. It also set-up a zookeeper connection watcher by default to re-set the data change watcher even during zk re-connect.
The monitor will set-up all watcher properly kick off the loop to process future event; you can also invoke loadZKData() in your class anytime to force reading zk data and apply it.
Example use cases are: - zookeeper based CachePoolCluster uses this to monitor cache pool members change - zookeeper based MigrationClient uses this ot monitor migration state transitioning
Cache specific cluster implementation.
Cache specific cluster implementation. - A cache pool is a Cluster of cache nodes. - cache pool requires a underlying pool manager as the source of the cache nodes - the underlying pool manager encapsulates logic of monitoring the cache node changes and deciding when to update the cache pool cluster
Cache pool config data object
ZooKeeper based cache pool cluster companion object
Cache pool config data format Currently this data format is only used by ZookeeperCachePoolManager to read the config data from zookeeper serverset parent node, and the expected cache pool size is the only attribute we need for now. In the future this can be extended for other config attributes like cache pool migrating state, backup cache servers list, or replication role, etc