Cassandra via-docker
Download as PPTX, PDF12 likes17,801 views
This document provides an overview of Apache Cassandra and how to deploy it using Docker. It begins with an introduction to Docker fundamentals and the CAP theorem. It then covers Cassandra fundamentals, including its history, users, benefits like scalability and consistency options, and challenges. The document demonstrates writes and reads in Cassandra with different consistency levels. It also discusses data replication strategies and calculating consistency. Finally, it previews a demo of deploying Cassandra with Docker and lists additional resources.
Cassandra via-docker
- 1. Apache Cassandra via Docker Chris Ballance Diligent Corporation @ballance
- 2. What we’ll cover Docker Fundamentals CAP Theorem Cassandra Fundamentals Demo of Cassandra deployed with Docker
- 4. Docker Hub “GIT for Virtual Machines” https://hub.docker.com/
- 5. Available Docker Images • Ubuntu, Centos, Debian, Fedora • Cassandra, mySQL, mongoDB • Node, Java, Erlang, Ruby, Rails • Wordpress • Redis • Hipache, NGINX • Create your own!
- 6. CAP THEOREM
- 7. Project Delivery Triangle Good Fast Cheap
- 8. Good Fast Cheap Project Delivery Triangle
- 9. Consistent Available Partition Tolerant CAP Theorem
- 10. CAP Theorem Consistent Available Partition Tolerant
- 14. Consistent Available Partition Tolerant Partition Tolerance is a myth!
- 15. We can guarantee Consistency, but not Availability. We may time out or fail to return anything Consistent Available Partition Tolerant
- 16. We can guarantee Availability, but our data may not be Consistent with other nodes Consistent Available Partition Tolerant
- 17. Our business needs will drive whether we choose Consistency or Availability Partition Tolerant
- 18. Implementations Consistent Available Partition TolerantAP Cassandra, CouchDB CP MongoDB, Big Table (GFS) CA RDBMS SQL Server
- 19. A contrived example for Consistency A B
- 20. User writes S1 to Node B A B S0 S0 S1
- 21. User queries Node A Nodes B & A have not sync’d A B S1S0
- 22. Query is blocked until B syncs with A A B S0 S1 ?
- 23. Once B syncs with A, the query on A is unblocked and returns S1 as expected A B S1 S1S1 S1 :-) * The query could potentially time out
- 24. A contrived example for Availability A B
- 25. User writes S1 to Node B A B S1 S0 S0
- 26. User queries Node A for S1 A B S1S0
- 27. Query returns current state of A, but is not consistent with B A B S1 S0 S0 S1 ?
- 28. A later query of A yields S1 previously written to B. Eventual consistency has been achieved. Idempotence Event Sourcing Local Caching Queueing A B S1 S1 S1 S1 :-)
- 29. Cassandra
- 30. History Developed for Facebook Inbox Search Created by one of the authors of Amazon Dynamo Released as Open Source in 2008 Became an Apache Incubator project in 2009 Graduated to an Apache Top Level Project
- 31. Who is using Cassandra in production today? Twitter Netflix Credit Suisse Cisco Many more… http://PlanetCassandra.org/companies/
- 32. Benefits Linear Scalability Data Sets can be larger than available memory Multi-master Built-in support for handling multiple data centers Decentralized & Distributed - No single point of failure Integrated caching Consistency options can be tuned through configuration Supports MapReduce Familiar query syntax - CQL (Cassandra Query Language) Designed for sparse loading of loosely typed data
- 34. Challenges JOINs are not supported! (not unique to Cassandra) Not for financial data (eventual consistency) Tooling is not yet as mature as Oracle or SQL Server
- 35. Why NOSQL?
- 36. Multiple persistence strategies Use the right tool for the job. Sometimes more than one tool is appropriate. NoSQL can work in conjunction with SQL solutions. For example, you might have trasactions you continue to store in a Relational Database, and build your new user social graph data in a NoSQL solution. Can’t decide on a local database for a mobile app? Sometimes the best route is to just persist the whole thing as JSON on disk until you need something faster.
- 37. Total Cost of Ownership (TCO) Let’s face it, SQL Server is expensive. If you need it, and can fully justify the cost, then it might be right for you. But as we’ve discussed, RDBMS can be a crutch and a default choice for persistence layers. Defaults are just that, default. They’re a catch-all that is rarely the best choice unless you’re resolving generic default problems
- 47. Write (consistency two) Client Coordinator Replicant Replicant Time-out Confirm Confirm Replicant
- 48. Write (consistency two) Client Coordinator Replicant Replicant Success Time-out Confirm Confirm Replicant
- 53. Read (consistency two) Client Coordinator Chosen Node Chosen Node
- 54. Read (consistency two) Client Coordinator Chosen Node Chosen Node
- 56. Read (consistency two) Client Coordinator Chosen Node Inconsistent Node
- 57. Read (consistency two) Client Coordinator Chosen Node Chosen Node Inconsistent Node
- 58. Read (consistency two) Client Coordinator Chosen Node Chosen Node Inconsistent Node
- 59. Read (consistency two) Client Coordinator Chosen Node Chosen Node Inconsistent Node Success
- 61. Consistency Modes ALL - Every node must have the data QUORUM - Most nodes must have the data ONE - At least one node must have the data TWO - At least two nodes must have the data THREE - At least three Nodes must have the data ANY - Any node has the data EACH_QUORUM - Each datacenter must have a quorum LOAL_QUORUM - Each node in the datacenter handling the request must have a quorum *Quorum = (replication_factor / 2 ) + 1
- 62. Calculating Consistency R + W > N R —> Read level consistency W —> Write level consistency N —> Number of replicas of the data
- 63. Data Replication SimpleStrategy - Single data center NetworkTopologyStrategy - Recommended strategy for multiple data centers. Provides Cassandra with info about the location of nodes by rack and datacenter
- 64. Deploying Cassandra with Docker Demo
- 65. Additional Resources Apache Cassandra http://cassandra.apache.org/ Docker https://www.docker.com/what-docker Docker Hub https://hub.docker.com/explore/ Cassandra for Developers - Paul O’Fallon https://www.pluralsight.com/courses/cassandra-developers DBA’s Guid to NoSQL: Apache Cassandra (Free eBook) http://is.gd/CassandraFreeEbook Cassandra 3.0 - DataStax PDF http://docs.datastax.com/en/cassandra/3.0/pdf/cassandra30.pdf
- 66. Questions?
Editor's Notes
- #7: Before I get into CAP theorem, I’d like to start off with another very similar triangle you’re already familiar with…
- #8: Before I get into CAP theorem, I’d like to start off with another very similar triangle you’re already familiar with…
- #10: CAP Theorem is similar. We have three competing interests
- #11: Choose two
- #12: Consistency - Always reads the most recent write - One version of the truth - But, if nodes are partitioned, we cannot guarantee consistency. - We can wait until we’re consistent, but in the meantime we may timeout and cannot continue.
- #13: Every (unfailed) node always returns query results Eventually consistent No errors or timeouts, but data may be stale If partitioned, a node can return its current state, but it may be an old version of the truth
- #14: Are all nodes always consistent? What is the lag time between node updates? One version of the truth Transactions are atomic
- #15: -Be dramatic… PARTITION TOLERANCE IS A MYTH!!1! -This is just the nature of networks. Nothing is guaranteed. -So really, we just have two competing interests, Consistency, and Availability. -Choosing which to favor cannot be an entirely technical decision, but something to discuss with your stakeholders and it depends on the type of app you’re building.
- #16: Once choice would be to favor Consistency. While this isn’t as popular with most workflows, it has its place.
- #17: -Another choice would be to favor Availability. You’ll find this is very common in social media, and media in general.
- #20: Here’s an example where we’re favoring consistency. We have a stick guy with data to write, and two nodes to write to and read from.
- #21: S1 is going to be our payload, and we send it to node B.
- #22: After we’ve sent our write (S1) to node B, we query Node A for the value of S. We might not even know that we’re being routed to node A for the read when we just wrote to node B. This can be even more confusing to even savvy users if this redirection is fully transparent behind a load balancer (and it typically is).
- #23: Stick man is not a patient fellow and he’s tapping his foot, waiting for his query to return. He’s probably just going to assume the system is broken after a few seconds…
- #24: Finally, Node A is able to synchronize with Node B to get S1 and return it to Stick guy. There’s a real risk of starving Stuck guy, and he’s already looking mighty thin.
- #25: Now let’s look at a different approach. This time, we will favor Availability over Consistency.
- #26: Our buddy, Stick guy, again sends S1 to Node B.
- #27: Now he queries Node A.
- #28: He immediately gets a reply, but it is stale data, since Nodes A & B have not yet synchronized. We don’t have to wait for the query to return, but we do get stale data that is older than what we have already sent out.
- #29: On the later query, stick guy gets back the value S1 that he is expecting, consistent with what he previously sent to Node B.