Module 10b - Apache ZooKeeper

Question 1

What is ZooKeeper?

Answer 1

ZooKeeper is a centralized service for maintaining configuration information, naming, providing distributed synchronization, and providing group services.

ZK is a high performance coordination service for distributed applications

Question 2

What does CAP stand for?

Answer 2

Consistency, Availability, Partition tolerance

Question 3

In a distributed system following a hierarchical structure, the master is prone to ______, and so we may want to spin up a ______ master. This is facilitated by a _______ service

Answer 3

failures
backup
coordination

Question 4

ZooKeeper’s architecture supports high ________ through ________ services. The clients can thus ask another ZooKeeper leader if the first fails to answer.

Answer 4

availability

redundant/replicated

Question 5

What protocols can you implement with ZooKeeper in a distributed system?

Answer 5

• Consensus
• Group Management
• Leader election
• Presence Protocols

Question 6

ZooKeeper exposes a simple set of ______ that ______ applications can build upon to implement ______ level services.

Answer 6

primitives
distributed
higher

Question 7

What high level services can the ZooKeeper primitives be used to implement?

Answer 7

• Synchronization
• Configuration Maintenance
• Groups
• Naming

Question 8

What are ZNodes?

Answer 8

Data registers. They are similar to files and directories.

Question 9

Filesystem files are designed for storage, whereas ZooKeeper data is kept _______, allowing Zookeeper to achieve high _______ and low _______.

Answer 9

in-memory
throughput
latency

Question 10

Like the distributed processes it coordinates, ZooKeeper itself is intended to be _______ over a set of hosts called an ensemble.

Answer 10

replicated

Question 11

The servers that make up the ZooKeeper service must all know about each other. As long as a ______ of the servers are available, the ZooKeeper service will be available

Answer 11

majority

Question 12

Clients connect to a single ZooKeeper server. The client maintains a ___ connection. If the TCP connection to the server breaks, the client will connect to a _____ server.

Answer 12

TCP

different

Question 13

ZooKeeper is fast in ______ dominant workloads

Answer 13

read

Question 14

The data stored at each ZNode is _____ in size

Answer 14

small

Question 15

ZNodes have a _______ number associated with them which is maintained with a stat structure. Each time a ZNode’s ______ is changed, this number is number increases

Answer 15

version

data

Question 16

The data stored at each ZNode in a namespace is read and written _____. A write to a ZNode _____ all the data

Answer 16

atomically

replaces

Question 17

Each ZNode has an ACL. What is an ACL?

Answer 17

Access Control List. Restricts who can do what

Question 18

ZooKeper has the concept of ephemeral nodes. What are ephemeral nodes?

Answer 18

They are Znodes that exists as long as the session that created that znode is active. When the session expires, the znode is deleted.

Question 19

Describe the concept of “watches” in ZooKeeper

Answer 19

Clients can set a “watch” on a znode. A watch will be triggered and removed when a znode changes. When a watch is triggered, the client receives a packet saying that the znode has changed.

Question 20

A guarantee that ZooKeeper provides is Sequential Consistency. Describe the term “sequential consistency” with respect to ZooKeeper

Answer 20

Updates from a client will be applied in the order that they were sent

Question 21

In ZooKeeper, what does atomicity mean with regards to updating data

Answer 21

Multiple clients updating a single resource will not encounter a race condition. An update will either succeed or fail. No partial Results

Question 22

What is it mean in ZooKeeper when we say it is a “Single System Image”?

Answer 22

A client will see the same view of the service regardless of the server that it connects to.

A client will never see an older view of the system even if the client fails over a different server with the same session.

Question 23

In ZooKeeper, once an ______ has been applied, it will ______ from that time forward until a client ______ the update

Answer 23

update
persist
overwrites

Question 24

We say that ZooKeeper is guaranteed to be “timeless” what does this mean?

Answer 24

The client’s view of the system is guaranteed to be up-to-date within a certain time bound

Question 25

Why is ZooKeeper very highly performant when reads outnumber writes?

Answer 25

Writes involve synchronizing the state of all servers - which has significant overhead

Question 26

In ZooKeeper, all write requests from clients are forwarded to a single server called the _____. The rest of the ZooKeeper servers are called ______

Answer 26

leader

followers

Question 27

What is the Curator API for ZooKeeper?

Answer 27

The Curator is a high-level API built on top of ZooKeeper’s official API. It handles the complexity of managing connections and provides a cleaner interface to the programmer

Question 28

What is the consistency model used by ZooKeeper?

Answer 28

Sequential consistency and linearizability (for writes)