SlideShare a Scribd company logo
1 of 11
Download to read offline
Copyright © 2020 Oracle and/or its affiliates.
MySQL NDB Cluster 8.0,
DBT2 Benchmark with 5TB Data
MySQL Cluster Development
Mikael Ronström
• 1 Data Node
• Intel Xeon Platinum 8260L @2.40 GHz
• 6 TB Intel Optane DC Persistent Memory
• 768 GB RAM
• Persistent Memory used in Memory mode
• DBT2 based on TPC-C specs with zero delay between
transactions
DBT2 Benchmark Definition
Persistent Memory Model in Benchmark
CPU
CPU Caches
DRAM
Persistent Memory
• Intel Xeon Platinum 8260L @2.40 GHz
768 GB DRAM
6 TB Intel Optane DC Persistent Memory
• Memory Mode => Memory cleared at restart, used as normal
memory
• App Direct Mode => Memory remains after restart
• In experiment here we used Memory Mode
Persistent Memory Models
DBT2 Benchmark Layout
DBT2 Driver
DBT2 Client
1 MySQL Server
1 NDB Data Node
Intel Xeon Platinum
8260L @2.40 GHz
2 sockets with
total of 48 cores
• Parallel LOAD DATA INFILE in 32 threads
• > 2 warehouses loaded per second
• 1 warehouse = 500.000 rows
• => More than 1 M Inserts per second
• 45.000 warehouses loaded in about 8 hours
• Number of warehouses limited by 5.9 TB SSD for REDO log and
checkpoint data, could load roughly 53.000 warehouses with a
larger SSD, will be verified later
DBT2 Load Phase
• DBT2 ran with 2 up to 512 threads. Default mode of DBT2 is
to use the warehouse id = thread id
• Default mode means all data accesses finds data in DRAM
cache (768 G in size)
• DBT2 altered mode means used warehouse is random and
thus benchmark will cause misses in DRAM cache that
requires reading and writing data in Persistent Memory layer
DBT2 Benchmark run
DBT2 Benchmark Results
TPM
0
100000
200000
300000
400000
2 8 16 32 64 128 192 256 384 512
Default Mode Altered Mode
• Optane memory increases transaction latency by 10-12% when
accessing the full range of 5 TB data
• Benchmark limited by MySQL Server, handles 370k TPM on a
single server with 5 TB user data
• NDB Cluster verified to handle properly DB sizes up to 5 TB on
one single 2 CPU Socket node
• With Optane DC Persistent Memory the recommendation is to
use hyperthreading also on LDM threads
DBT2 5 TB Conclusions
• MySQL Server used 1 socket + 3 cores
• Benchmark used 2 cores
• NDB Data Node used 19 cores (24 LDM threads on 12 cores)
• Base latency was around 7 ms for NewOrder transactions
(fairly complex transactions, lots of SQL statements)
Benchmark Configuration
Thank you!

More Related Content

What's hot

TokuDB internals / Лесин Владислав (Percona)
TokuDB internals / Лесин Владислав (Percona)TokuDB internals / Лесин Владислав (Percona)
TokuDB internals / Лесин Владислав (Percona)
Ontico
 
Availability and scalability in mongo
Availability and scalability in mongoAvailability and scalability in mongo
Availability and scalability in mongo
Md. Khairul Anam
 
S3332 peter bakkum
S3332 peter bakkumS3332 peter bakkum
S3332 peter bakkum
Peter Bakkum
 

What's hot (20)

TokuDB internals / Лесин Владислав (Percona)
TokuDB internals / Лесин Владислав (Percona)TokuDB internals / Лесин Владислав (Percona)
TokuDB internals / Лесин Владислав (Percona)
 
NUMA and Java Databases
NUMA and Java DatabasesNUMA and Java Databases
NUMA and Java Databases
 
OpenTSDB: HBaseCon2017
OpenTSDB: HBaseCon2017OpenTSDB: HBaseCon2017
OpenTSDB: HBaseCon2017
 
Exploring Parallel Merging In GPU Based Systems Using CUDA C.
Exploring Parallel Merging In GPU Based Systems Using CUDA C.Exploring Parallel Merging In GPU Based Systems Using CUDA C.
Exploring Parallel Merging In GPU Based Systems Using CUDA C.
 
Elasticsearch avoiding hotspots
Elasticsearch  avoiding hotspotsElasticsearch  avoiding hotspots
Elasticsearch avoiding hotspots
 
Tuning Linux for Databases.
Tuning Linux for Databases.Tuning Linux for Databases.
Tuning Linux for Databases.
 
Extended memory access in PHP
Extended memory access in PHPExtended memory access in PHP
Extended memory access in PHP
 
Availability and scalability in mongo
Availability and scalability in mongoAvailability and scalability in mongo
Availability and scalability in mongo
 
Tales from production with postgreSQL at scale
Tales from production with postgreSQL at scaleTales from production with postgreSQL at scale
Tales from production with postgreSQL at scale
 
Garbage collection in JVM
Garbage collection in JVMGarbage collection in JVM
Garbage collection in JVM
 
An Introduction to Apache Cassandra
An Introduction to Apache CassandraAn Introduction to Apache Cassandra
An Introduction to Apache Cassandra
 
Logs @ OVHcloud
Logs @ OVHcloudLogs @ OVHcloud
Logs @ OVHcloud
 
Pgxc scalability pg_open2012
Pgxc scalability pg_open2012Pgxc scalability pg_open2012
Pgxc scalability pg_open2012
 
Jvm & Garbage collection tuning for low latencies application
Jvm & Garbage collection tuning for low latencies applicationJvm & Garbage collection tuning for low latencies application
Jvm & Garbage collection tuning for low latencies application
 
S3332 peter bakkum
S3332 peter bakkumS3332 peter bakkum
S3332 peter bakkum
 
Memcached
MemcachedMemcached
Memcached
 
Bluestore
BluestoreBluestore
Bluestore
 
RAIDZ on-disk format vs. small blocks
RAIDZ on-disk format vs. small blocksRAIDZ on-disk format vs. small blocks
RAIDZ on-disk format vs. small blocks
 
Jvm tuning for low latency application & Cassandra
Jvm tuning for low latency application & CassandraJvm tuning for low latency application & Cassandra
Jvm tuning for low latency application & Cassandra
 
Understanding DLmalloc
Understanding DLmallocUnderstanding DLmalloc
Understanding DLmalloc
 

Similar to Ndb cluster 80_dbt2_5_tb

Oracle Exadata Version 2
Oracle Exadata Version 2Oracle Exadata Version 2
Oracle Exadata Version 2
Jarod Wang
 
Colvin exadata mistakes_ioug_2014
Colvin exadata mistakes_ioug_2014Colvin exadata mistakes_ioug_2014
Colvin exadata mistakes_ioug_2014
marvin herrera
 
Exaflop In 2018 Hardware
Exaflop In 2018   HardwareExaflop In 2018   Hardware
Exaflop In 2018 Hardware
Jacob Wu
 

Similar to Ndb cluster 80_dbt2_5_tb (20)

MySQL NDB Cluster 8.0 SQL faster than NoSQL
MySQL NDB Cluster 8.0 SQL faster than NoSQL MySQL NDB Cluster 8.0 SQL faster than NoSQL
MySQL NDB Cluster 8.0 SQL faster than NoSQL
 
Ndb cluster 80_oci_dbt2
Ndb cluster 80_oci_dbt2Ndb cluster 80_oci_dbt2
Ndb cluster 80_oci_dbt2
 
August 2013 HUG: Removing the NameNode's memory limitation
August 2013 HUG: Removing the NameNode's memory limitation August 2013 HUG: Removing the NameNode's memory limitation
August 2013 HUG: Removing the NameNode's memory limitation
 
cachegrand: A Take on High Performance Caching
cachegrand: A Take on High Performance Cachingcachegrand: A Take on High Performance Caching
cachegrand: A Take on High Performance Caching
 
Real world repairs
Real world repairsReal world repairs
Real world repairs
 
Collier exadata technical overview presentation 4 14-10
Collier exadata technical overview presentation 4 14-10Collier exadata technical overview presentation 4 14-10
Collier exadata technical overview presentation 4 14-10
 
Theta and the Future of Accelerator Programming
Theta and the Future of Accelerator ProgrammingTheta and the Future of Accelerator Programming
Theta and the Future of Accelerator Programming
 
Argonne's Theta Supercomputer Architecture
Argonne's Theta Supercomputer ArchitectureArgonne's Theta Supercomputer Architecture
Argonne's Theta Supercomputer Architecture
 
Oracle Exadata Version 2
Oracle Exadata Version 2Oracle Exadata Version 2
Oracle Exadata Version 2
 
Accelerating hbase with nvme and bucket cache
Accelerating hbase with nvme and bucket cacheAccelerating hbase with nvme and bucket cache
Accelerating hbase with nvme and bucket cache
 
Colvin exadata mistakes_ioug_2014
Colvin exadata mistakes_ioug_2014Colvin exadata mistakes_ioug_2014
Colvin exadata mistakes_ioug_2014
 
High Performance Hardware for Data Analysis
High Performance Hardware for Data AnalysisHigh Performance Hardware for Data Analysis
High Performance Hardware for Data Analysis
 
Mike Pittaro - High Performance Hardware for Data Analysis
Mike Pittaro - High Performance Hardware for Data Analysis Mike Pittaro - High Performance Hardware for Data Analysis
Mike Pittaro - High Performance Hardware for Data Analysis
 
Exaflop In 2018 Hardware
Exaflop In 2018   HardwareExaflop In 2018   Hardware
Exaflop In 2018 Hardware
 
Percona FT / TokuDB
Percona FT / TokuDBPercona FT / TokuDB
Percona FT / TokuDB
 
Accelerating HBase with NVMe and Bucket Cache
Accelerating HBase with NVMe and Bucket CacheAccelerating HBase with NVMe and Bucket Cache
Accelerating HBase with NVMe and Bucket Cache
 
Exadata_X10M-Hardware-Overview.pdf
Exadata_X10M-Hardware-Overview.pdfExadata_X10M-Hardware-Overview.pdf
Exadata_X10M-Hardware-Overview.pdf
 
MariaDB ColumnStore
MariaDB ColumnStoreMariaDB ColumnStore
MariaDB ColumnStore
 
Cassandra TK 2014 - Large Nodes
Cassandra TK 2014 - Large NodesCassandra TK 2014 - Large Nodes
Cassandra TK 2014 - Large Nodes
 
TiDB vs Aurora.pdf
TiDB vs Aurora.pdfTiDB vs Aurora.pdf
TiDB vs Aurora.pdf
 

Recently uploaded

The title is not connected to what is inside
The title is not connected to what is insideThe title is not connected to what is inside
The title is not connected to what is inside
shinachiaurasa2
 
Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024
Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024
Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024
VictoriaMetrics
 
%+27788225528 love spells in Boston Psychic Readings, Attraction spells,Bring...
%+27788225528 love spells in Boston Psychic Readings, Attraction spells,Bring...%+27788225528 love spells in Boston Psychic Readings, Attraction spells,Bring...
%+27788225528 love spells in Boston Psychic Readings, Attraction spells,Bring...
masabamasaba
 

Recently uploaded (20)

WSO2CON2024 - It's time to go Platformless
WSO2CON2024 - It's time to go PlatformlessWSO2CON2024 - It's time to go Platformless
WSO2CON2024 - It's time to go Platformless
 
%in tembisa+277-882-255-28 abortion pills for sale in tembisa
%in tembisa+277-882-255-28 abortion pills for sale in tembisa%in tembisa+277-882-255-28 abortion pills for sale in tembisa
%in tembisa+277-882-255-28 abortion pills for sale in tembisa
 
Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...
Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...
Devoxx UK 2024 - Going serverless with Quarkus, GraalVM native images and AWS...
 
Right Money Management App For Your Financial Goals
Right Money Management App For Your Financial GoalsRight Money Management App For Your Financial Goals
Right Money Management App For Your Financial Goals
 
The title is not connected to what is inside
The title is not connected to what is insideThe title is not connected to what is inside
The title is not connected to what is inside
 
call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️
call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️
call girls in Vaishali (Ghaziabad) 🔝 >༒8448380779 🔝 genuine Escort Service 🔝✔️✔️
 
8257 interfacing 2 in microprocessor for btech students
8257 interfacing 2 in microprocessor for btech students8257 interfacing 2 in microprocessor for btech students
8257 interfacing 2 in microprocessor for btech students
 
WSO2CON 2024 - Does Open Source Still Matter?
WSO2CON 2024 - Does Open Source Still Matter?WSO2CON 2024 - Does Open Source Still Matter?
WSO2CON 2024 - Does Open Source Still Matter?
 
WSO2Con2024 - Enabling Transactional System's Exponential Growth With Simplicity
WSO2Con2024 - Enabling Transactional System's Exponential Growth With SimplicityWSO2Con2024 - Enabling Transactional System's Exponential Growth With Simplicity
WSO2Con2024 - Enabling Transactional System's Exponential Growth With Simplicity
 
WSO2CON 2024 - Cloud Native Middleware: Domain-Driven Design, Cell-Based Arch...
WSO2CON 2024 - Cloud Native Middleware: Domain-Driven Design, Cell-Based Arch...WSO2CON 2024 - Cloud Native Middleware: Domain-Driven Design, Cell-Based Arch...
WSO2CON 2024 - Cloud Native Middleware: Domain-Driven Design, Cell-Based Arch...
 
%in kempton park+277-882-255-28 abortion pills for sale in kempton park
%in kempton park+277-882-255-28 abortion pills for sale in kempton park %in kempton park+277-882-255-28 abortion pills for sale in kempton park
%in kempton park+277-882-255-28 abortion pills for sale in kempton park
 
Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024
Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024
Large-scale Logging Made Easy: Meetup at Deutsche Bank 2024
 
Crypto Cloud Review - How To Earn Up To $500 Per DAY Of Bitcoin 100% On AutoP...
Crypto Cloud Review - How To Earn Up To $500 Per DAY Of Bitcoin 100% On AutoP...Crypto Cloud Review - How To Earn Up To $500 Per DAY Of Bitcoin 100% On AutoP...
Crypto Cloud Review - How To Earn Up To $500 Per DAY Of Bitcoin 100% On AutoP...
 
%+27788225528 love spells in Boston Psychic Readings, Attraction spells,Bring...
%+27788225528 love spells in Boston Psychic Readings, Attraction spells,Bring...%+27788225528 love spells in Boston Psychic Readings, Attraction spells,Bring...
%+27788225528 love spells in Boston Psychic Readings, Attraction spells,Bring...
 
Payment Gateway Testing Simplified_ A Step-by-Step Guide for Beginners.pdf
Payment Gateway Testing Simplified_ A Step-by-Step Guide for Beginners.pdfPayment Gateway Testing Simplified_ A Step-by-Step Guide for Beginners.pdf
Payment Gateway Testing Simplified_ A Step-by-Step Guide for Beginners.pdf
 
MarTech Trend 2024 Book : Marketing Technology Trends (2024 Edition) How Data...
MarTech Trend 2024 Book : Marketing Technology Trends (2024 Edition) How Data...MarTech Trend 2024 Book : Marketing Technology Trends (2024 Edition) How Data...
MarTech Trend 2024 Book : Marketing Technology Trends (2024 Edition) How Data...
 
Direct Style Effect Systems - The Print[A] Example - A Comprehension Aid
Direct Style Effect Systems -The Print[A] Example- A Comprehension AidDirect Style Effect Systems -The Print[A] Example- A Comprehension Aid
Direct Style Effect Systems - The Print[A] Example - A Comprehension Aid
 
%in Hazyview+277-882-255-28 abortion pills for sale in Hazyview
%in Hazyview+277-882-255-28 abortion pills for sale in Hazyview%in Hazyview+277-882-255-28 abortion pills for sale in Hazyview
%in Hazyview+277-882-255-28 abortion pills for sale in Hazyview
 
Software Quality Assurance Interview Questions
Software Quality Assurance Interview QuestionsSoftware Quality Assurance Interview Questions
Software Quality Assurance Interview Questions
 
%in Bahrain+277-882-255-28 abortion pills for sale in Bahrain
%in Bahrain+277-882-255-28 abortion pills for sale in Bahrain%in Bahrain+277-882-255-28 abortion pills for sale in Bahrain
%in Bahrain+277-882-255-28 abortion pills for sale in Bahrain
 

Ndb cluster 80_dbt2_5_tb

  • 1. Copyright © 2020 Oracle and/or its affiliates. MySQL NDB Cluster 8.0, DBT2 Benchmark with 5TB Data MySQL Cluster Development Mikael Ronström
  • 2. • 1 Data Node • Intel Xeon Platinum 8260L @2.40 GHz • 6 TB Intel Optane DC Persistent Memory • 768 GB RAM • Persistent Memory used in Memory mode • DBT2 based on TPC-C specs with zero delay between transactions DBT2 Benchmark Definition
  • 3. Persistent Memory Model in Benchmark CPU CPU Caches DRAM Persistent Memory • Intel Xeon Platinum 8260L @2.40 GHz 768 GB DRAM 6 TB Intel Optane DC Persistent Memory
  • 4. • Memory Mode => Memory cleared at restart, used as normal memory • App Direct Mode => Memory remains after restart • In experiment here we used Memory Mode Persistent Memory Models
  • 5. DBT2 Benchmark Layout DBT2 Driver DBT2 Client 1 MySQL Server 1 NDB Data Node Intel Xeon Platinum 8260L @2.40 GHz 2 sockets with total of 48 cores
  • 6. • Parallel LOAD DATA INFILE in 32 threads • > 2 warehouses loaded per second • 1 warehouse = 500.000 rows • => More than 1 M Inserts per second • 45.000 warehouses loaded in about 8 hours • Number of warehouses limited by 5.9 TB SSD for REDO log and checkpoint data, could load roughly 53.000 warehouses with a larger SSD, will be verified later DBT2 Load Phase
  • 7. • DBT2 ran with 2 up to 512 threads. Default mode of DBT2 is to use the warehouse id = thread id • Default mode means all data accesses finds data in DRAM cache (768 G in size) • DBT2 altered mode means used warehouse is random and thus benchmark will cause misses in DRAM cache that requires reading and writing data in Persistent Memory layer DBT2 Benchmark run
  • 8. DBT2 Benchmark Results TPM 0 100000 200000 300000 400000 2 8 16 32 64 128 192 256 384 512 Default Mode Altered Mode
  • 9. • Optane memory increases transaction latency by 10-12% when accessing the full range of 5 TB data • Benchmark limited by MySQL Server, handles 370k TPM on a single server with 5 TB user data • NDB Cluster verified to handle properly DB sizes up to 5 TB on one single 2 CPU Socket node • With Optane DC Persistent Memory the recommendation is to use hyperthreading also on LDM threads DBT2 5 TB Conclusions
  • 10. • MySQL Server used 1 socket + 3 cores • Benchmark used 2 cores • NDB Data Node used 19 cores (24 LDM threads on 12 cores) • Base latency was around 7 ms for NewOrder transactions (fairly complex transactions, lots of SQL statements) Benchmark Configuration