You signed in with another tab or window.Reload to refresh your session.You signed out in another tab or window.Reload to refresh your session.You switched accounts on another tab or window.Reload to refresh your session.Dismiss alert
Attempt to fix unstable Result Cache regression tests
force_parallel_mode = regress is causing a few more problems than Ithought. It seems that both the leader and the single worker cancontribute to the execution. I had mistakenly thought that only the workerprocess would do any work. Since it's not deterministic as to whichof the two processes will get a chance to work on the plan, it seems justbetter to disable force_parallel_mode for these tests. At least doingthis seems better than changing to EXPLAIN only rather than EXPLAINANALYZE.Additionally, I overlooked the fact that the number of executions of thesub-plan below a Result Cache will execute a varying number of timesdepending on cache eviction. 32-bit machines will use less memory andevict fewer tuples from the cache. That results in the subnode beingexecuted fewer times on 32-bit machines. Let's just blank out the numberof loops in each node.