Commit 63c80fc
committed
tests: avoid use of loopback dialer in Benchmark{Parallel,}Sysbench
When follow-the-workload is disabled (#153865), leases no longer
reliably stay on n1. Since the workload clients are always
created on n1, they were able to use the loopback dialer (which
avoids going through the TCP layer by using inter-progress
communication instead) prior to that change.
We want these benchmarks to be independent of lease placement,
so we disable the loopback dialer. That this was not done
initially is likely an oversight.
This commit will cause a a benchmark regression, but it is
due to changing what is being benchmarked. The regression
does not result in any production change.1 parent 52aaf2e commit 63c80fc
2 files changed
+14
-13
lines changed| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
89 | 89 | | |
90 | 90 | | |
91 | 91 | | |
| 92 | + | |
92 | 93 | | |
93 | 94 | | |
94 | 95 | | |
| |||
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
27 | 27 | | |
28 | 28 | | |
29 | 29 | | |
| 30 | + | |
30 | 31 | | |
| 32 | + | |
31 | 33 | | |
32 | 34 | | |
33 | 35 | | |
| |||
164 | 166 | | |
165 | 167 | | |
166 | 168 | | |
| 169 | + | |
| 170 | + | |
| 171 | + | |
| 172 | + | |
| 173 | + | |
| 174 | + | |
| 175 | + | |
| 176 | + | |
| 177 | + | |
| 178 | + | |
167 | 179 | | |
168 | 180 | | |
169 | 181 | | |
170 | 182 | | |
171 | | - | |
172 | | - | |
173 | | - | |
174 | | - | |
175 | | - | |
| 183 | + | |
176 | 184 | | |
177 | 185 | | |
178 | 186 | | |
179 | 187 | | |
180 | 188 | | |
181 | 189 | | |
182 | | - | |
183 | | - | |
184 | | - | |
185 | | - | |
186 | 190 | | |
187 | 191 | | |
188 | 192 | | |
| |||
376 | 380 | | |
377 | 381 | | |
378 | 382 | | |
379 | | - | |
380 | | - | |
381 | | - | |
382 | | - | |
383 | 383 | | |
384 | 384 | | |
385 | 385 | | |
| |||
0 commit comments