Mentor Graphics Customer Presentation
Download as PPTX, PDF2 likes1,596 views
Mentor Graphics used Splunk to gain insights into their storage and client infrastructure. They collected data from over 100 filers and 8,500 systems to generate reports on capacity, performance, errors and changes over time. This helped optimize resources, track issues and answer questions from various divisions. Splunk provided a cost-effective solution compared to vendor tools and helped enhance visibility, monitoring and reporting across their global systems.
Mentor Graphics Customer Presentation
- 1. Copyright © 2015 Splunk Inc. Lee Melvin & Shawn Roberts Mentor Graphics
- 2. Hello my name is: Lee Melvin Mentor Graphics Technical Architect - Storage Hello my name is: Shawn Roberts Mentor Graphics Technical Architect - Clients
- 3. Mentor Graphics® provides software and hardware design solutions that enable companies to develop better electronic products faster and more cost- effectively.
- 4. How I got started with Splunk • Dual vendor network-attached storage strategy • 100+ filers across 30+ sites, used by 10+ divisions • Needed ways to • Provide per-division summary reporting, accessible to engineers on demand • Implement performance service level objectives • Provide storage administrators with actionable issues, not just graphs and metrics • Provide reporting that spans vendors • Vendor reporting tools expensive and inadequate for these needs • Previously, layered a fragile and manual scripting process on top • Initial focus on capacity and performance analysis Lee
- 5. Our “Storage” Solution Forwarder/Data Collection Node Lee Custom Vendor Specific Add-on Splunk’s plugin Indexer/search head
- 6. How I got started with Splunk • We had ~8,500 systems around the world that were not being monitored • No visibility into what could be causing occasional crashes • No ability to: • Add another agent • “increase the IT tax” • Negatively impact performance • Track changes over time • People were asking questions… Shawn
- 9. Shawn
- 10. Shawn
- 11. Shawn WMI? Event Logs? Event Forwarding? 8,500 Systems Performance hit to network and clients 8,500 Systems 8,500 Systems! Event Viewer <> Reporting
- 12. Shawn
- 13. Shawn
- 14. Shawn
- 16. Doesn’t add another agent Doesn’t “increase the IT tax” No impact to performance Track changes over time Shawn
- 17. Tough client reports now possible Boot Time Average aka What’s normal? Last 30 vs 365 Top/Bottom 10 Error Events Top 10 Errors by category Startup impacting Application crashes ! Shawn
- 18. Shawn
- 19. Storage reporting enhanced! • Tagging filesystems by division, product group, performance tier, or other abstract (via Splunk lookup table) • Custom technology add-on to ingest data from Hitachi HNAS • Capacity/Performance reports that span both NetApp and Hitachi • Division summary and per-filesystem views • Actionable items for storage administrators • Outliers view – slow, busy, near-full • Stale data reporting – big chunks of unused data Lee
- 20. Lee
- 21. Lee
- 22. Lee
- 24. No good deed, etc…what’s next? • Lookup/compare by: • Location • AD OU • Manager • Type • Model • Manufacturer Shawn
- 25. Client dashboard • Measurements • Make, Model, CPU, Memory, Disks, Encryption status, etc. • Age of Windows installation • Most common errors • Last N errors • SCCM Data • Installed software • Update status • Changes! • Hardware • Software • AV Data • Versions • Activity Shawn
- 26. Client dashboard with Splunk • Measurements • Make, Model, CPU, Memory, Disks, Encryption status, etc. • Age of Windows installation • Most common errors • Last N errors • SCCM Data • Installed software • Update status • Changes! • Hardware • Software • AV Data • Versions • ActivityX vs Normal X over Time Shawn
- 27. Storage reporting • Reports • Filer configuration auditing • Geographic visualization • Long term (5-year) reports leveraging summary indexes • Matching workloads to appropriate performance tier • Alerting • Alert filesystem owners on near- capacity using lookup table • Infrastructure • Web frontend to edit tags in lookup table • Division-specific view roles to limit access/confusion • Ingest additional HDS metrics • Ingest Symantec NetBackup logs • Round the learning curve on data models, simplify Lee
- 28. Lessons learned • You have data in more places than you might think. • You don’t always need a huge license purchase to answer hard questions. • Use the provided apps for inspiration (and for cut/paste). Don’t treat them as an end state. Shawn Lee
- 29. Copyright © 2015 Splunk Inc. Thank You! Shawn Roberts shawn_roberts@mentor.com +ShawnRobertsNerd Lee Melvin lee_melvin@mentor.com linkedin.com/in/lmelvin
Editor's Notes
- #8: Mostly Windows OS
- #9: And BTW, they are clients.
- #10: I went shopping for a “Client Monitoring” tool. There aren’t any options in this space.
- #11: We have a great working relationship with Microsoft. I asked them for guidance.
- #12: Here’s our conversation.
- #13: Now that I could get the logs in one place, it was much easier to go shopping.
- #14: I happened to be attending OSCON around the same time…
- #15: I talked with Splunk.
- #16: With Windows Event Forwarding and Splunk, I’m able to put together a fairly simple solution.
- #17: All green.
- #19: Also, we are able to perform on demand reporting to solve tough client problems like this one.
- #24: We’re a little bruised from the journey, but we’re in a good place.
- #26: The next project I’ve been tasked with is creating a Client Dashboard to help our help desk and desktop administrators. Without Splunk, I can do a lot with DB queries and HTML.
- #27: With Splunk I can call out what’s important in all that data. This is really useful to our admins who don’t have a lot of familiarity with each system they work with.