Data Center Automation (from Velocity 2017)

Transcript

1. DATA CENTER AUTOMATION DAVID RADCLIFFE David Radcliffe @dwradcliffe
3. @dwradcliffe David Radcliffe OUR HISTORY OUR PLAN OUR SOLUTION OUR AUTOMATION
4. @dwradcliffe David Radcliffe SPREADSHEETS WRONG SPREADSHEETS !! A FEW YEARS AGO…
5. @dwradcliffe David Radcliffe COLLINS BUT STILL NO AUTOMATION ! LAST YEAR…
6. @dwradcliffe David Radcliffe
7. @dwradcliffe David Radcliffe % NOW…
8. @dwradcliffe David Radcliffe TRACK EVERYTHING
9. @dwradcliffe David Radcliffe ACCURATELY TRACK EVERYTHING Inform our supply chain Automate our provisioning
10. @dwradcliffe David Radcliffe DATA WE COLLECT FROM SERVERS ‣ hostname ‣ os details ‣ product information and serial numbers ‣ chef status ‣ ipmi status and address ‣ network interfaces and network neighbors ‣ disks ‣ cpus ‣ memory
11. @dwradcliffe David Radcliffe DATA CENTER ROOM CAGE RACK CABINET CHASSIS SERVER CPU MEMORY DISK NIC
12. @dwradcliffe David Radcliffe DATA CENTER GENESIS SERVER AGENT SERVER AGENT SERVER AGENT
13. @dwradcliffe David Radcliffe
14. @dwradcliffe David Radcliffe
15. @dwradcliffe David Radcliffe
16. @dwradcliffe David Radcliffe
17. @dwradcliffe David Radcliffe
18. @dwradcliffe David Radcliffe
19. @dwradcliffe David Radcliffe REPORTS ‣ Hierarchy and Location Audits ‣ Problem areas ‣ Unclassified servers ‣ Devices with no location ‣ Unhealthy devices ‣ Missing/Duplicate IPMI ‣ Capacity ‣ Automation ‣ Spare disks/memory
20. @dwradcliffe David Radcliffe AUTOMATION ‣ Genesis acts as a queue ‣ Remote Actions are queued up and automation agents will run them ‣ Move all our playbooks into scripts with a friendly UI, logging, authentication, etc. ‣ Safety!
21. @dwradcliffe David Radcliffe BOOTSTRAP PIPELINE CHEF SERVERSPEC TEST KITCHEN LINECOOK DEPENDABLE REPEATABLE SERVER IMAGES
22. @dwradcliffe David Radcliffe BOOTSTRAP PIPELINE GENESIS AGENT ZYGOTE PXE BOOT LIVESYS CHEF REBOOT CHEF DONE DISK SETUP
23. @dwradcliffe David Radcliffe
24. @dwradcliffe David Radcliffe REMOTE HANDS IPAD ACCESS TO GENESIS AUDIT USING REPORTS PHOTOS DEEP LINKS TO GENESIS
25. @dwradcliffe David Radcliffe OTHER INTEGRATIONS ‣ DNS ‣ Mute node alerts during maintenance actions ‣ Sync device metadata with monitoring tools ‣ Kubernetes
26. @dwradcliffe David Radcliffe TAKEAWAYS
27. @dwradcliffe David Radcliffe THANK YOU!