SF Big Analytics_20190612: Scaling Apache Spark on Kubernetes at Lyft
Download as PPTX, PDF1 like342 views
Li Gao from Lyft discusses their initiatives for scaling Apache Spark on Kubernetes, focusing on the data infrastructure landscape, challenges faced, and the advantages of using Kubernetes. The presentation highlights the ongoing transition from vendor-based solutions to Spark on Kubernetes, emphasizing scalability and resource management. Despite significant progress, challenges remain, such as scaling limits and dependency management.
SF Big Analytics_20190612: Scaling Apache Spark on Kubernetes at Lyft
- 2. Introduction 2#UnifiedAnalytics #SparkAISummit Li Gao Works in the Data Platform team at Lyft, currently leading multiple Data Infra initiatives within Data Platform, including the Spark on Kubernetes project. Previously held tech leadership roles at Salesforce, Fitbit, Groupon, and other startups.
- 3. Agenda 3#UnifiedAnalytics #SparkAISummit ● Introduction of Data Landscape at Lyft ● The challenges we face ● How Apache Spark on Kubernetes can help ● Remaining work
- 4. Data Landscape 4#UnifiedAnalytics #SparkAISummit ● Batch data Ingestion and ETL ● Data Streaming ● ML platforms ● Notebooks and BI tools ● Query and Visualization ● Operational Analytics ● Data Discovery & Lineage ● ML workflow orchestration ● Cloud Platforms
- 5. Business Analysts, Data Scientists, AML Cloud Infra Services Data Infrastructure @ Lyft 5 External and Internal Products and Services Data Portal (Discovery, WF, SLx dashboard etc. Data Infra
- 6. Batch Compute Clusters Batch Compute @ Lyft 6 Events Ext Data RDB/KV Sys Events IngestPipelines AWSS3 AWSS3 HMS Presto,HiveClient,andBITools Analysts Engineers Scientists Services
- 7. Evolving Batch Architecture 7 Future2016-2017 Vendor-based Hadoop 2017-2018 Hive on MR Vendor Presto Mid 2018 Hive on Tez + Spark Adhoc Late 2018 Spark on Vendor GA Early-Mid 2019 Spark on K8s Alpha Spark on K8s Beta & Preprod
- 9. Batch Compute Challenges 9 ● 3rd Party vendor dependency limitations ● Data ETL expressed solely in SQL, and sometimes in hard-maintain complex SQL ● Complex logic expressed in Python that hard to adopt in SQL ● Different dependencies and versions ● Resource load balancing for heterogeneous workloads
- 10. Is SQL the complete solution? 10
- 11. How Spark can help? 11 RDB/KV Application s APIs Environments Data Sources and Data Sinks
- 12. What challenges remain? 12 ● Per job custom dependencies and security context isolation ● Multi version runtime requirements (Py3 v.s. Py2, Spark versions) ● Still need to run on shared clusters for cost efficiency ● Mixed ML and ETL workloads
- 13. How Kubernetes can help? 13 Operators & Controllers Pods Ingress Services Namespaces Pods Immutability Event driven & Declarative Vibrant CNCF Community ServiceMesh Multi-TenancySupport Image Registry
- 15. What challenges still remain? ● Spark on k8s is still in its early days ● Single cluster scaling limit ● CRD operator choking limit ● Cluster control plane rollout pain points ● Pod churn and IP allocations throttling ● Default k8s scheduler limit ● ECR container registry reliability 15
- 16. Current scale 16 ● 10s PB data lake ● (O) 100k batch jobs running daily ● ~ 1000s of EC2 nodes spanning multiple clusters and AZs ● ~ 1000s of workflows running daily
- 17. How Lyft scales Spark on K8s # of Clusters # of Namespaces # of Pods Pod Churn Rate # of Nodes Pod Size Job:Pod ratio IP Alloc Rate Limit ECR Rate Limit Affinity & Isolation QoS & Quota Pod Scheduler
- 19. A Start of a Spark Job @ k8s 19 Resource Labels Job Cluster Pool Cluster Namespace Group Namespace Spark CRD K8s Pods ● (1) and (2) Dispatcher Gateway ● (3) Cluster Controller ● (4) Job Scheduler ● (5) Namespace Group Controller ● (6) Spark Operator ● (7) K8s Pod Scheduler (1) (3) (4) (5) (6) (7) (2)
- 21. HA in Cluster Pool 21 Cluster 1 Cluster 2 Cluster 3 Cluster Pool A Cluster 4 ● Cluster rotation within a cluster pool ● Automated provisioning of a new cluster and (manually) add into rotation ● Throttle at lower bound when rotation in progress
- 22. Multiple Namespaces (Groups) 22 Pod Pod Pod Namespace 1 Pod Pod Pod Namespace 2 Pod Pod Pod Namespace 3 Node A Node B Node C Node D Role1 Role1 Role2 Max Pod Size 1 Max Pod Size 2 ● Practical ~3K active pods per namespace observed ● Less preemption required when namespace isolated by quota ● Different namespaces can map different IAM roles and sidecar configurations for security and auditing purposes
- 23. Pod Sharing 23 Job Controller Spark Driver Pod Spark Exec Pods Job 2 Driver Pod Job 2 Exec Pods Job 3 Driver Pod Job 3 Exec Pods Shared Pods Job 1 Job 4 Job 3 Job 2 AWS S3 Dep Dep Dedicate & Isolated Pods Dep
- 24. Separate DML from DDL 24
- 25. DDL Separation to reduce churn 25
- 26. Pod Priority and Preemption (WIP) 26 ● Priority base preemption ● Driver pod has higher priority than executor pod ● Experimental D1 D2 E1 E2 E3 E4 K8s Scheduler D1 E5 New Pod Req Before D2 E5 E2 E3 E4 After E1 Evictedhttps://github.com/kubernetes/kubernetes/issues/71486 https://github.com/kubernetes/enhancements/issues/564
- 27. Taints and Tolerations (WIP) 27 Node A Node B Node C Node D Node E Node F P1 P2 P3 P4 P5 P6 P7 P7 P8 P9 P10 Controllers and Watchers Job 1 Job 2 Core Nodes (Taint) Worker Nodes (Taint) ● Other considerations: Node Labels, Node Selectors to separate GPU and CPU based workloads
- 28. Mutating Admission Hooks 28 K8S API HTTP Handler Authn & Authz Mutating admin controllers Schema validation validating admin controllers ETCD k8s pod scheduler kubelet Node Spark Pod Mutating admin webhooks validating admin webhooks Pod Request kubelet Node Spark Pod sidecars config credit: https://banzaicloud.com/blog/k8s-admission-webhooks/
- 29. Custom k8s Pod scheduler for batch (WIP) Predicates Priorities Round Robin Predicates Weight Engine Placement Engine Policies Default k8s scheduler Dynamic Policy Driven k8s scheduler All Active Notes All Active Notes
- 30. What about ECR reliability? 30 Node 1 Node 2 Node 3 Pods Pods Pods DaemonSet + Docker In Docker ECR Container Images
- 31. Spark Job Config Overlays (DML) 31 Cluster Pool Defaults Cluster Defaults Spark Job User Specified Config Cluster and Namespace Overrides Final Spark Job Config Config Composer & Event Watcher Spark Operator
- 32. X-Ray of Job Controller 32
- 33. Controllers & Watchers • Job scheduler • Spark job config composer • Namespace group controller • k8s pod scheduler • Service controllers (STS, Jupyter) • K8s metrics & events watchers • Spark job/crd events & metrics watchers 33
- 34. X-Ray of Spark Operator 34
- 35. Monitoring and Logging Toolbox 35 JMX
- 36. Provision & Automation 36 Kustomize Template K8S Deploy Sidecar injectors Secrets injectors DaemonSets KIAM
- 37. Remaining work ● More intelligent & efficient job routing, scheduler and parameter composer ● End-to-End serverless, self-serviceable, and user- oriented data compute mesh ● Fine grained cost attribution ● Improved docker image distribution ● Spark 3.0 & Kubernetes v1.14+ 37
- 38. Key Takeaways ● Apache Spark can help unify different batch data compute use cases ● Kubernetes can help solve the dependency and multi-version requirements using its containerized approach ● Spark on Kubernetes can scale significantly by using a multi- cluster compute mesh approach with proper resource isolation and scheduling techniques ● Challenges remain when running Spark on Kubernetes at scale 38
- 39. Community 39 This effort would not be possible without the help from the open source and wider communities:
- 41. 41
- 42. 42
- 43. Monitoring Example - OOM Kill 43
- 44. What about dependencies? 44 RTree Libraries Data CodecsSpatial Libraries
- 45. 3rd Party Vendor Limitations 45 ● Proprietary patches ● Inconsistent bootstrap ● Release schedule ● Homogeneous environments
- 46. What about Python functions? 46 “I want to express my processing logic in python functions with external geo libraries (i.e. Geomesa) and interact with Hive tables” --- Lyft data engineer
Editor's Notes
- #5: Different users and usecases - ml, streaming, realtime , batch, notebooks multiple cloud platforms
- #14: declarative predictable & repeatable operators add extensibility multi tenancy container nati ve
- #15: CNCF is a vibrant community and supports numerous projects
- #16: patch rollout/updates for crd/control plane is still evolving pod churn - etcd/resource/ttl/ip allocation in ec2 for eg
- #46: What is homogeneous envs here?