Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
UrbonBayesPere_BerlinBuzzwords18_ConnectingDataInfraWithDataFlow.pdf
Search
Pere Urbón
June 13, 2018
0
130
UrbonBayesPere_BerlinBuzzwords18_ConnectingDataInfraWithDataFlow.pdf
Pere Urbón
June 13, 2018
Tweet
Share
More Decks by Pere Urbón
See All by Pere Urbón
Building a self-service Kafka Platform
purbon
0
290
10 ways to deploy Apache Kafka® and have fun along the way
purbon
1
170
Apache Kafka: advice from the trenches or how to successfully fail!
purbon
3
1.1k
Building an Streaming Platform with Kafka
purbon
0
94
Connecting the data infrastructure with the DataFlow (Apache NiFi)
purbon
2
220
Ten Tips to completely fail building your Search Project
purbon
0
150
Learning to Rank 101, Bringing personalisation to data discovery
purbon
0
460
The quantum mechanics of data pipelines
purbon
0
100
Data Engineering without borders
purbon
0
71
Featured
See All Featured
The Language of Interfaces
destraynor
151
23k
It's Worth the Effort
3n
180
27k
Statistics for Hackers
jakevdp
789
220k
The Success of Rails: Ensuring Growth for the Next 100 Years
eileencodes
30
6k
Practical Orchestrator
shlominoach
182
9.7k
How to Ace a Technical Interview
jacobian
272
22k
Fantastic passwords and where to find them - at NoRuKo
philnash
37
2.5k
Code Review Best Practice
trishagee
55
15k
Docker and Python
trallard
34
2.7k
Designing Dashboards & Data Visualisations in Web Apps
destraynor
226
51k
Cheating the UX When There Is Nothing More to Optimize - PixelPioneers
stephaniewalter
274
13k
RailsConf & Balkan Ruby 2019: The Past, Present, and Future of Rails at GitHub
eileencodes
125
32k
Transcript
@purbon Connecting the data infrastructure with the DataFlow
@purbon Pere Urbon-Bayes Software Architect pere.urbon@{gmail.com, acm.org}
Topics for Today • Integration patterns for the enterprise startup.
• What is Apache NIFI. • Examples • NiFi on operation (best practises).
@purbon Integrate all the things!
@purbon Enterprise integration is the task of making separate applications
work together to produce an unified set of functionality. The applications probably run on multiple computers, which may be geographically dispersed.
@purbon Some application might need to be integrated even though
they were not designed for integration and can not be changed. This issues, and others, are what makes application integration difficult.
@purbon Each integration faces different needs and criteria, we can
group them as Application coupling Integration simplicity Data formats and timeliness Data or functionality Communication
@purbon There is only a limited set of integration options
@purbon File transfer
@purbon Shared database
@purbon RPC invoke
@purbon Messaging
@purbon Enterprise Integration Patterns
@purbon What is Apache NiFi?
@purbon An easy to use, powerful, and reliable system to
process and distribute data. Web-based interface Highly configurable Data Provenance Designed for extension Secure
@purbon NiFi was build to automate the flow of data
between systems. an automated and managed flow of information between systems. But what is Dataflow?
@purbon How Apache NiFi look like
@purbon Concepts behind Apache NiFi
@purbon A Flow file
@purbon The Flow file Processor
@purbon A Connection
@purbon A Process Group
@purbon Apache NiFi Architecture Distributed using Apache Zookeper
@purbon Let’s take a closer look…
@purbon Apache NiFI Operations
@purbon Maximum file handles hard nofile 50000 soft nofile 50000
/etc/security/limits.conf
@purbon Maximum forked Procs hard nproc 10000 soft nproc 10000
/etc/security/limits.conf /etc/security/limits.d/90-nproc.conf
@purbon Increase number of TCP sockets sudo sysctl -w net.ipv4.ip_local_port_range="10000
65000"
@purbon Timeout sockets in TIMED_WAIT state sudo sysctl -w net.ipv4.netfilter.ip_conntrack_tcp_timeout_time_wait="1"
@purbon Never SWAP vm.swappiness = 0 /etc/sysctl.conf /dev/sda7 /chroot ext2
defaults, noatime 1 2 /etc/fstab
@purbon Thanks a lot! Questions? disagreements? threads? Pere Urbon-Bayes Data
Wrangler pere.urbon@{gmail.com, acm.org}