Downloaded 44 times
Uploaded byMongoDB
MongoDB Days UK: Run MongoDB on Google Cloud Platform
This document discusses MongoDB performance on Google Cloud Platform. It provides benchmarks comparing MongoDB performance on Google Compute Engine virtual machines with different disk configurations. The benchmarks show that dedicating separate disks for the MongoDB database files and journal files significantly improves write performance. The document also describes how the company uses MongoDB on Google Cloud Platform for time-series database workloads, including off-site backups to Google Cloud Storage and automated restore testing.
More Related Content
Similar to MongoDB Days UK: Run MongoDB on Google Cloud Platform
![MongoDB .local San Francisco 2020: Powering the new age data demands [Infosys]](https://cdn.slidesharecdn.com/ss_thumbnails/315pminfosysfinalsfoversionvocalpart1-200120221508-thumbnail.jpg?width=640&height=640&fit=bounds)
MongoDB Days UK: Run MongoDB on Google Cloud Platform
- 1. Google confidential │Do not distribute Google confidential │ Do not distribute MongoDB on Google Cloud Platform Tom Grey EMEA Head of Cloud Platform Solutions Engineering at Google Jorge Salamero A Technical Overview
- 2. Everybody is talkingabout cloud... Google Confidential & Proprietary
- 3. Making Google lookeasy is hard... Google Confidential & Proprietary
- 4. Powering google.com... Google Confidential& Proprietary Google is the fourth largest server manufacturer in the world after Dell, HP, and IBM according to Martin Reynolds of Garter Group
- 5. Powering google.com... Google Confidential& Proprietary If Google were an ISP it would be the second largest ISP by traffic on the planet according to Arbor Networks
- 6. Powering google.com... Google Confidential& Proprietary Customised hardware built from cheap commodity parts Software resilience and easy repair not hardware resilience Horizontal layers not vertical stacks Vast numbers of homogeneous servers managed at scale
- 7. Google confidential │Do not distribute Google confidential │ Do not distribute Externalising Google Omega Application Runtimes & Services ● Iterate & deploy fast ● Scale to global demand ● Standards compliant Data Services ● Data Intelligence ● Designed for Big Data ● High Performance Map Reduce Dremel Pregel Percolator Data Storage and Distribution ● Global Resilient Architecture ● Global Edge Distribution ● Huge Secure Capacity GFS Colossus Spanner BigTable F1 Global Data Centre & Networks ● Highly Resilient, Efficient & Performant ● 3rd Largest Server Manufacturer ● 2nd Largest Global Data Network Google Research Publications referenced are available here: http://research.google.com/pubs/papers.html Google Products etc... App Engine Cloud Endpoints BigQuery Cloud Storage Cloud SQL Cloud DataStore Compute Engine Your Applications Externalising Google
- 8. Google confidential │Do not distribute Google confidential │ Do not distribute Storage Compute Ops Big Data Network The Google Cloud Platform
- 9. Google confidential │Do not distribute Google confidential │ Do not distribute The Google Cloud Platform Google Cloud Storage Google Cloud SQL Google Cloud Datastore Google Cloud Bigtable Compute Ops Big Data Network
- 10. Google confidential │Do not distribute Google confidential │ Do not distribute The Google Cloud Platform Google Cloud Storage Google Cloud SQL Google Cloud Datastore Google Cloud Bigtable Google Compute Engine Google App Engine Google Container Engine Ops Big Data Network
- 11. Google confidential │Do not distribute Google confidential │ Do not distribute The Google Cloud Platform Google Cloud Storage Google Cloud SQL Google Cloud Datastore Google Cloud Bigtable Google Compute Engine Google App Engine Google Container Engine Google BigQuery Google Cloud Dataflow Google Cloud Datalab Google Cloud Pub/Sub Google Cloud Dataproc Ops Network
- 12. Google confidential │Do not distribute Google confidential │ Do not distribute The Google Cloud Platform Google Cloud Storage Google Cloud SQL Google Cloud Datastore Google Cloud Bigtable Google Compute Engine Google App Engine Google Container Engine Google BigQuery Google Cloud Dataflow Google Cloud Datalab Google Cloud Pub/Sub Google Cloud Dataproc Google Cloud Networking Ops
- 13. Google confidential │Do not distribute Google confidential │ Do not distribute The Google Cloud Platform Google Cloud Storage Google Cloud SQL Google Cloud Datastore Google Cloud Bigtable Google Compute Engine Google App Engine Google Container Engine Google BigQuery Google Cloud Dataflow Google Cloud Datalab Google Cloud Pub/Sub Google Cloud Networking Google Cloud Monitoring Google Cloud Logging Google Cloud Dataproc
- 14. Google confidential │Do not distribute Google confidential │ Do not distribute The Google Cloud Platform Google Cloud Storage Google Cloud SQL Google Cloud Datastore Google Cloud Bigtable Google Compute Engine Google App Engine Google Container Engine Google BigQuery Google Cloud Dataflow Google Cloud Datalab Google Cloud Pub/Sub Google Cloud Networking Google Cloud Monitoring Google Cloud Logging Google Cloud Dataproc
- 15. Google confidential │Do not distribute Google confidential │ Do not distribute API Project Anatomy of a Compute Engine project
- 16. Google confidential │Do not distribute Google confidential │ Do not distribute CLI UI Code API Project Anatomy of a Compute Engine project
- 17. Google confidential │Do not distribute Google confidential │ Do not distribute CLI UI Code API VM VM VM Project Anatomy of a Compute Engine project
- 18. Google confidential │Do not distribute Google confidential │ Do not distribute CLI UI Code API Persistent Disk Cloud Storage VM VM VM Project Anatomy of a Compute Engine project
- 19. Google confidential │Do not distribute Google confidential │ Do not distribute Private Network CLI UI Code API Persistent Disk Cloud Storage VM VM VM Project Anatomy of a Compute Engine project
- 20. Google confidential │Do not distribute Google confidential │ Do not distribute Internet Private Network CLI UI Code API Persistent Disk Cloud Storage VM VM VM Project Anatomy of a Compute Engine project
- 21. Google confidential │Do not distribute Google confidential │ Do not distribute Region Regions: Geographic location of resources Inter-Region Latency > Zone Group Latency Zones: Independent of other Zones Isolated within a Region Distribute instances across Zones to protect against single zone system failure.
- 22. Google confidential │Do not distribute Google confidential │ Do not distribute Machine Types
- 23. Google Confidential andProprietary MongoDB on Google Compute Engine + our experience @
- 24. Google Confidential andProprietary
- 25. Google Confidential andProprietary cloud infrastructure monitoring Monitoring as-a-Service servers & services 30+ integration custom plugins + API dashboard + alerts
- 26. Google Confidential andProprietary How we use MongoDB? Time series database Since 2009 250+TB/month 6000 writes/sec 500,000,000 new documents per day Ubuntu 12.04 LTS Bare metal servers SSD disks Puppet Forge MongoDB module
- 27. Google Confidential andProprietary MongoDB on Cloud Servers Traditionally MongoDB on VM issues CPU steal from other guests (no high CPU requirements itself) Disk IO Google Compute intelligent throttling no more noisy neighbours provisioned IOPS predefined instances types
- 28. Google Confidential andProprietary Google Compute Engine Disks Standard Persistent Disk (storage backed by hard disk drives) SSD PD (storage backed by solid state drives) LOCAL SSD (not persistent, obviously) Standard: sustained performance increases with size + burst for peaks 100 GB: 30 random read IOPS or 150 random write IOPS (12 MB/s for reads and 9 MB/s for writes) 10 TB: 3000 random read IOPS or 15000 random write IOPS (180 MB/s for reads and 120 MB/s for writes) ● SSD: IOPS increase faster, throughput the same max 10000 random read IOPS at 333 GB max 15000 random write IOPs at 500 GB ● Network egress cap: redundancy 3.3 x IOPS (2Gbps / CPU)
- 29. Google Confidential andProprietary Google Compute Engine Disks: example SATA2 7200RPM ~75 IOPS / 120MB/s IO Pattern SSD PD size required 75 random reads 250 GB 75 random writes 50 GB 120 MB/s read 1000 GB 120 MB/s writes 1333 GB
- 30. Google Confidential andProprietary Google Compute Engine Disks: local SSD 375 GB, up to 4 per VM many limitations: no redundancy, no snapshots, create-time only NVMe or SCSI Standard PD SSD PD Local SSD NVMe Read IOPS per GB 0.3 30 453.3 Write IOPS per GB 1.5 30 240 Read IOPS per instance 3,000 10,000 680,000 Write IOPS per instance 15,000 15,000 360,000
- 31. Google Confidential andProprietary So, remember In Google Compute, IOPS scale linearly with volume size
- 32. Google Confidential andProprietary Dimensioning and tuning FS: DICARD/TRIM, lazy init (Google takes care) Low readahead (blockdev) IO queue depth: 1 each 400-800 IOPS max 64 ↑ depth ↑ IOPS but ↑ latency 1 CPU each 2000 read IOPS / 2500 write IOPS
- 33. Google Confidential andProprietary Recommended configuration dbpath separated volume journal on a different volume likely to be big volume for IOPS, low usage ~ at least 200GB ~6000 write IOPS cannot use snapshots for backups (fsync lock or shutdown required) directoryperdb for each db (locking is managed at database level) flexibility, performance independent
- 34. Google Confidential andProprietary Our tests 3 different performance scenarios: 1. no extra disks (default 10GB volume) 2. dedicated dbpath (200GB volume) 3. dedicated dbpath (200GB volume) + dedicated journal (200GB volume) 2 different nodes: 1. n1-standard-2 (2 vCPUs and 7.5GB RAM) 2. n1-highmem-8 (8 vCPUs and 52GB RAM)
- 35. Google Confidential andProprietary The results on n1-standard-2
- 36. Google Confidential andProprietary The results on n1-highmem-8
- 37. Google Confidential andProprietary Conclusions No much difference until you start to acknowledge writes Difference between instances is small, real-life test needed Validated the recommendations, again: separate your dbpath separate your journal Full details: https://blog.serverdensity.com/mongodb-on-google-compute-engine-tips-and-benchmarks/
- 38. Google Confidential andProprietary How we use MongoDB on Google Compute? MongoDB Cloud Manager with backup verification in GC real-time offsite backups (only a few seconds behind) replica node for each replica set copy of every write operation sustained traffic 42Mbps point-in-time restores
- 39. Google Confidential andProprietary How we use MongoDB on Google Compute? Off-site backups into GC API to trigger restore job: get a tarball Downloaded to Google Cloud storage versioning included regional buckets for redundancy USA + EU
- 40. Google Confidential andProprietary How we use MongoDB on Google Compute? Restore on MongoDB GC Launch instance with SSD PD gsutil to download the tarball untar the backup install MongoDB ● COMPARE with PRODUCTION ● whole process: ~10 min, twice a day ● Python, Buildbot, notifications on HipChat
- 41. Google Confidential andProprietary Thanks!
- 42. Q&A Confidential | Donot distribute