Exercise 1: Why the Slowdown and the Crashes Shared a Root Cause — Possible Solution ==================================================================== HOW THE TWO SYMPTOMS WERE ORIGINALLY REPORTED ------------------------------ The afternoon slowdown and the database client crashes were reported as two separate, unrelated problems - a general performance complaint and a specific application-crash complaint, with no obvious connection between them at the outset. WHAT STEP 1's INVESTIGATION ACTUALLY FOUND ------------------------------ Per this chapter, Resource Monitor's Disk tab, "sorted by IOPS rather than raw throughput, identifies the database client itself issuing a huge volume of tiny read requests." The disk activity behind the afternoon slowdown wasn't background system activity or an unrelated process - it was the same database client application, already struggling, generating the disk load. WHAT STEP 2's INVESTIGATION REVEALED ABOUT THE CRASHES ------------------------------ Per this chapter, Event Viewer showed "repeated Error-level entries... describing a failed read against a local cache file," ultimately traced to "a corrupted value under the client's own HKCU branch." The database client was repeatedly trying, and failing, to read its own corrupted local cache - each failed attempt generating disk activity (the many small reads Step 1 identified) before eventually crashing outright. WHY THIS MEANS THE TWO SYMPTOMS SHARE ONE ROOT CAUSE ------------------------------ The corrupted Registry value causing the client's cache-read failures is the single underlying cause behind both symptoms: the repeated, failing read attempts generated the disk I/O responsible for the afternoon slowdown, and continuing to fail eventually caused the application to crash outright. What looked like two separate problems was actually one malfunctioning process producing two different visible effects, depending on how long the failing behavior continued before the application gave up entirely. WHY THIS IS A GENUINELY IMPORTANT LESSON FOR MULTI-SYMPTOM DIAGNOSIS ------------------------------ Investigating each reported symptom using the appropriate tool from earlier in this course, rather than assuming from the outset that separately reported symptoms must have separate causes, is exactly what allowed this shared root cause to surface - treating them as independent from the start could have led to two separate, unnecessary investigations chasing what turned out to be the same underlying problem. WHY THIS WORKS AS AN ANSWER ------------------------------ It traces the specific evidence connecting both symptoms back to the same corrupted Registry value and the client's own repeated failed cache reads, and explains why investigating each symptom on its own terms - rather than assuming unrelated causes - is what revealed the shared root cause.