Keep the vote cache processor out of the benchmarks The fixture waited for the vote cache processor's queue to drain after starting elections, but the processor swaps the queue out before it looks the hashes up, so the wait could end while the lookups were still running and overlapping the first votes of a measurement. The cache is empty whenever elections start here, so there is nothing to replay: the processor is now disabled like the schedulers, and the wait is gone.
__________
Type: commit
Hash: a6bb69d70d9b95d27e65b439c029fa0cc71d5d63
Created: 21 day(s) ago
__________
gr0v1ty/nano-node:a6bb69d70d9b95d27e65b439c029fa0cc71d5d63
| Test Case | Status | Duration | Median Deviation | Log | Graph |
|---|---|---|---|---|---|
| 5n4pr_conf_10k_bintree | PASS | 111 s | -5.0 s | GitHub job logs | Show |
| 5n4pr_conf_10k_change | PASS | 156 s | -7.0 s | GitHub job logs | Show |
| 5n4pr_conf_change_dependant | PASS | 116 s | 1.5 s | GitHub job logs | - |
| 5n4pr_conf_change_independant | PASS | 113 s | 2.5 s | GitHub job logs | - |
| 5n4pr_conf_send_dependant | PASS | 118 s | 2.0 s | GitHub job logs | - |
| 5n4pr_conf_send_independant | PASS | 115 s | 1.5 s | GitHub job logs | - |
| 5n4pr_rocks_10k_bintree | PASS | 117 s | 5.5 s | GitHub job logs | - |
| 5n4pr_rocks_10k_change | PASS | 184 s | 9.5 s | GitHub job logs | - |