Hadoop Security. Building a fence around your Hadoop cluster. Lars Francke June 12, Berlin Buzzwords 2017

Transcription

1 Hadoop Security Building a fence around your Hadoop cluster Lars Francke June 12, 2017 Berlin Buzzwords 2017

2 Introduction

3 About me - Lars Francke Partner & Co-Founder at OpenCore Before that: EMEA Hadoop Consultant Hadoop since 2008/2009 Apache Committer: Hive, ORC Contact: 2017 OpenCore GmbH & Co. KG 1/50

4 Overview

5 Overview - Be Warned? 2017 OpenCore GmbH & Co. KG 2/50

6 The Book Source: oreilly.com 2017 OpenCore GmbH & Co. KG 3/50

7 Other Books Source: oreilly.com 2017 OpenCore GmbH & Co. KG 4/50

8 Other Sources Kerberos: The Definitive Guide 1 Kerberos (German) 2 Active Directory, 5th Edition 3 Hadoop and Kerberos: The Madness beyond the Gate 4 HBase: The Definitive Guide, 2nd Edition 5 Bulletproof SSL and TLS OpenCore GmbH & Co. KG 5/50

9 Project Management How is a typical project structured and where does Hadoop security come into play? 2017 OpenCore GmbH & Co. KG 6/50

10 Project Management Idea/Initiation 2017 OpenCore GmbH & Co. KG 7/50

11 Project Management Idea/Initiation Planning/Design 2017 OpenCore GmbH & Co. KG 7/50

12 Project Management Idea/Initiation Planning/Design Execution/Implementation 2017 OpenCore GmbH & Co. KG 7/50

13 Project Management Idea/Initiation Planning/Design Execution/Implementation Production 2017 OpenCore GmbH & Co. KG 7/50

14 When To Start You can not begin thinking about Security too early! 2017 OpenCore GmbH & Co. KG 8/50

15 Next steps Let s dig into the details of each step in a project 2017 OpenCore GmbH & Co. KG 9/50

16 Project Planning

17 What To Think About? What am I going to use the cluster for? 2017 OpenCore GmbH & Co. KG 10/50

18 What To Think About? What am I going to use the cluster for? Which tools am I going to need to achieve the use-cases? 2017 OpenCore GmbH & Co. KG 10/50

19 What To Think About? What am I going to use the cluster for? Which tools am I going to need to achieve the use-cases? What kind of data am I going to ingest, store or process? 2017 OpenCore GmbH & Co. KG 10/50

20 What To Think About? What am I going to use the cluster for? Which tools am I going to need to achieve the use-cases? What kind of data am I going to ingest, store or process? What kind of corporate guidelines exist that must be followed? 2017 OpenCore GmbH & Co. KG 10/50

21 Do Not Assume Anything Do not assume companies/people know what a product is capable of just because they seem confident OpenCore GmbH & Co. KG 11/50

22 What does Security even mean? What is part of Hadoop Security? 2017 OpenCore GmbH & Co. KG 12/50

23 Parts Of A Security Concept Authentication of users 2017 OpenCore GmbH & Co. KG 13/50

24 Parts Of A Security Concept Authentication of users Authorization of users 2017 OpenCore GmbH & Co. KG 13/50

25 Parts Of A Security Concept Authentication of users Authorization of users Auditing 2017 OpenCore GmbH & Co. KG 13/50

26 Parts Of A Security Concept Authentication of users Authorization of users Auditing Data Protection: Encryption on-the-wire 2017 OpenCore GmbH & Co. KG 13/50

27 Parts Of A Security Concept Authentication of users Authorization of users Auditing Data Protection: Encryption on-the-wire Data Protection: Encryption at-rest 2017 OpenCore GmbH & Co. KG 13/50

28 Project Execution

29 Kerberos == Hadoop Security? So, does Hadoop Security mean I need to enable Kerberos and I m done? 2017 OpenCore GmbH & Co. KG 14/50

30 Kerberos == Hadoop Security? So, does Hadoop Security mean I need to enable Kerberos and I m done? No! Far from it! 2017 OpenCore GmbH & Co. KG 14/50

31 Overview Before you re able to secure anything 2017 OpenCore GmbH & Co. KG 15/50

32 Overview Before you re able to secure anything something must be running that can be secured OpenCore GmbH & Co. KG 15/50

33 Overview - contd OpenCore GmbH & Co. KG 16/50

34 Network 2017 OpenCore GmbH & Co. KG 17/50

35 Host/Operating System SELinux NTP/Chrony Firewalls Antivirus Proxies 2017 OpenCore GmbH & Co. KG 18/50

36 Hadoop Installation You usually install a cluster manager first 2017 OpenCore GmbH & Co. KG 19/50

37 Hadoop Installation You usually install a cluster manager first Then you install Agents 2017 OpenCore GmbH & Co. KG 19/50

38 Hadoop Installation You usually install a cluster manager first Then you install Agents Followed by a distribution with lots of components 2017 OpenCore GmbH & Co. KG 19/50

39 Authentication Core Hadoop requires Kerberos for strong authentication Multiple choices on how to implement that: Cluster-local standalone KDC Cluster-local standalone KDC with one-way cross-realm trust to a central KDC Direct integration into a central KDC 2017 OpenCore GmbH & Co. KG 20/50

40 Authentication: Cloudera Manager/Ambari Wizards Source: unsplash.com by Sirotorn Sumpunkulpak 2017 OpenCore GmbH & Co. KG 21/50

41 Authentication - contd. What you need: Names/IPs for your KDC Supported Encryption Types (see my blog post 7 for more details) Firewall must allow all cluster machines to access the KDC Potentially a bunch of information to configure krb5.conf properly OpenCore GmbH & Co. KG 22/50

42 Authentication - contd. Ideally you want the automatic option, but... You need an account in your AD/KDC that is allowed to create other accounts! (AD only) You need to talk via LDAPS and make sure that you have all the truststores set up properly (AD only) You cannot have multiple SPNs per User For SPNEGO you need a HTTP/<host> principal, sometimes doesn t match corporate guidelines 2017 OpenCore GmbH & Co. KG 23/50

43 Authentication - contd. There s a manual option 2017 OpenCore GmbH & Co. KG 24/50

44 Authentication - contd. Kerberos is only part of the authentication story OpenCore GmbH & Co. KG 25/50

45 Authentication - contd. Knox, Cloudera Manager, Ambari, Ranger, Sentry, Hive, Impala etc. all have their own authentication layers which also support LDAP(S) or other mechanisms OpenCore GmbH & Co. KG 26/50

46 Identity Management Your users need to exist on the workers so that YARN can start jobs using setuid() OpenCore GmbH & Co. KG 27/50

47 Identity Management - contd. This is usually done using a tool like SSSD (free) or Centrify (commercial) 2017 OpenCore GmbH & Co. KG 28/50

48 Identity Management - contd. This is usually done using a tool like SSSD (free) or Centrify (commercial) When using Centrify, configure it to not create the default HTTP principal for each server 2017 OpenCore GmbH & Co. KG 28/50

49 Identity Management - contd. This is usually done using a tool like SSSD (free) or Centrify (commercial) When using Centrify, configure it to not create the default HTTP principal for each server This does not mean that your users need to be able to SSH into the workers, quite the opposite 2017 OpenCore GmbH & Co. KG 28/50

50 Identity Management - contd. You need to have the details needed to fetch user & group information from LDAP(S) Some solutions require a domain join 2017 OpenCore GmbH & Co. KG 29/50

51 Impersonation/Proxy Users 2017 OpenCore GmbH & Co. KG 30/50

52 Authorization All tools have some form of authorization built-in Core Hadoop components have a first line defense: Service Level Authorizations Ranger & Sentry promise cross-cutting RBAC functionality 2017 OpenCore GmbH & Co. KG 31/50

53 Data Protection: Encryption At-Rest HDFS Transparent Encryption All I need is a KMS, right? 2017 OpenCore GmbH & Co. KG 32/50

54 Data Protection: Encryption At-Rest - contd. KMS should be close to your clients and NameNode KMS should be separately administered AFAIK only Hadoop 3 will allow the / (root) directory to be its own zone and have sub-zones Secure access to the backing store & host where KMS runs You can use a HSM What about Impersonation and things like Knox or HttpFS? 2017 OpenCore GmbH & Co. KG 33/50

55 Data Protection: Encryption At Rest - contd. Now that my data in HDFS is secure I m good, right? Metadata in databases Loq & Audit files YARN localized stuff Temporary files (Spark) spill files & cached data Only Cloudera has a (paid) solution for this: Cloudera NavEncrypt 2017 OpenCore GmbH & Co. KG 34/50

56 Data Protection/Authorization: Data Masking Static vs. dynamic masking Ranger & BigSQL support masking, Sentry does not RecordService might be a potential integration point 3rd party tools 2017 OpenCore GmbH & Co. KG 35/50

57 Data Protection: Encryption On-The-Wire Web Interfaces REST/Thrift Interfaces MapReduce Shuffle Spark Shuffle (only in 2.1+) Server to Agent communication Ingest/Egress tools (Flume, Informatica, Kafka, ) RPC HDFS Data traffic Traffic from your YARN apps 2017 OpenCore GmbH & Co. KG 36/50

58 Data Protection: Encryption On-The-Wire 2-way TLS possible? Which cipher suites are supported? Some tools automatically disable HTTP others don t! Missing documentation all around 2017 OpenCore GmbH & Co. KG 37/50

59 Auditing & Logging No documentation (on auditing events and lots of other things) Ranger aggregates audit logs, async by default Cloudera has Navigator No integrity protection Need to protect against admins 2017 OpenCore GmbH & Co. KG 38/50

60 3rd Party Tools Not a single product has good documentation How do they access the cluster? Do they themselves authenticate and authorize users? How? Auditing/Logging? 2017 OpenCore GmbH & Co. KG 39/50

61 Production

62 What now? So you ve finished the project and handed it over to the operations team what now? 2017 OpenCore GmbH & Co. KG 40/50

63 Keep it running Plan regular updates to your OS Use a Configuration Management tool (e.g. Ansible) Plan regular updates to your distribution! 2017 OpenCore GmbH & Co. KG 41/50

64 Monitoring All our logging & auditing is irrelevant if you re not monitoring & alerting on those logs OpenCore GmbH & Co. KG 42/50

65 Environments You need a place where you can test upgrades. You might also need a place for backups. Devs would like a cluster as well 2017 OpenCore GmbH & Co. KG 43/50

66 Environments 2017 OpenCore GmbH & Co. KG 44/50

67 Environments 2017 OpenCore GmbH & Co. KG 45/50

68 User Lifecycle What happens when a user leaves your organization? Caches Ranger Usersync Existing sessions Shared accounts 2017 OpenCore GmbH & Co. KG 46/50

69 Misc

70 Cloud? What about the cloud providers? 2017 OpenCore GmbH & Co. KG 47/50

71 Distributions Cloudera 2017 OpenCore GmbH & Co. KG 48/50

72 Distributions Cloudera Hortonworks 2017 OpenCore GmbH & Co. KG 48/50

73 Distributions Cloudera Hortonworks IBM 2017 OpenCore GmbH & Co. KG 48/50

74 Distributions Cloudera Hortonworks IBM Microsoft HDInsight 2017 OpenCore GmbH & Co. KG 48/50

75 Thank You! Thank you for listening! 2017 OpenCore GmbH & Co. KG 49/50

76 Questions & Contact Questions? Contact me Visit us at opencore.com And if this stuff interests you: We re looking to expand our team! 2017 OpenCore GmbH & Co. KG 50/50