Value Based Testing

1,073 views
923 views

Published on

0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
1,073
On SlideShare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
30
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Value Based Testing

  1. 1. ODTUGValue-Based Testing Your mileage may vary, void where prohibited 1
  2. 2. ODTUGWikipedia.com:Quality assurance (QA) refers to the planned andsystematic activities implemented in a quality systemso that quality requirements for a product or servicewill be fulfilled. It is the systematic measurement,comparison with a standard, monitoring of processesand an associated feedback loop that confers errorprevention. Your mileage may vary, void where prohibited 2
  3. 3. ODTUGWhat do they really do? • Ensure proper functionality • Look for “big picture” issues • Large-scale testing (hopefully) • Identify edge cases • Advocate for user experience • Environment is volatile Your mileage may vary, void where prohibited 3
  4. 4. What do they really do? Old school process: • Throw code over the wallDEV QA OPS • And run • Its Ops problem now Your mileage may vary, void where prohibited 4
  5. 5. ODTUG What do they really do?DEV • Responsibility lines blurred • Collaboration required QA • Dev should be in escalation pathOPS Your mileage may vary, void where prohibited 5
  6. 6. ODTUG Ops changes also need QA• New tools, scripts, jobs must follow release procedures• Patches, PSUs, environment settings• Can combine to test all at once• Monitoring must be tested (verified) too Your mileage may vary, void where prohibited 6
  7. 7. ODTUG But how?• Make identical environments (i.e. not a laptop VM!)• Automation ensures repeatable results, fewer mistakes• Two-way communications – Ops takes input from QA – QA takes input from Ops Your mileage may vary, void where prohibited 7
  8. 8. ODTUG Conducting Tests…Better• QA needs *real* info about production• Session counts, workload stats, response time• Apply load identical to prod load.• When identical isn’t possible, scale back, note risks Your mileage may vary, void where prohibited 8
  9. 9. ODTUG More Tests• Operational tests – HA failures, backups, stats gathering• Avg load, peak load, crash load• Load test with missing nodes, disks, NICs Your mileage may vary, void where prohibited 9
  10. 10. ODTUGWhat did you do? • Record tests completed, especially parameters • Keep performance reports, AWR, capture ASH data, key exec plans • Baselines = key, especially around upgrades Your mileage may vary, void where prohibited 10
  11. 11. ODTUG Typical Challenges, Issues• Spend all the money on prod, no $$$ for QA• Our staff is so skilled, we don’t need a “real” QA• Our schedule slipped, so we cut QA time down to 10% of the original plan Your mileage may vary, void where prohibited 11
  12. 12. ODTUG Typical Challenges, Issues• We don’t have requirements• Can’t you test it faster?• We don’t have a way to simulate 500 users Your mileage may vary, void where prohibited 12
  13. 13. ODTUG Answer to Typical Challenges• Me: What is the cost of downtime?• Mgmt: We aren’t sure…• Me: Let me take production down, then we can measure.• Mgmt: On second thought, I think I can probably estimate it pretty closely without an outage. Your mileage may vary, void where prohibited 13
  14. 14. ODTUGWar Stories Your mileage may vary, void where prohibited
  15. 15. ODTUG Q&A• Any questions or comments, please contact: – Gwen Shapira • Email: cshapi@gmail.com • Twitter: @gwenshap – Dan Norris • Email: dannorris@dannorris.com • Twitter: @dannorris Your mileage may vary, void where prohibited 15
  16. 16. Your mileage may vary, void where prohibited

×