Expire deferred cementing entries while the loop is idle (#5292) * Comment the cleanup step of the cementing set loop * Expire deferred cementing entries while the loop is idle The cementing loop waited without a timeout when nothing was queued, and only evicts deferred entries at the top of each pass. An entry that failed to cement stayed deferred past its age cutoff until new work arrived, so cementing_failed did not fire and the hash kept counting as being confirmed. The idle wait now wakes up every 5 seconds, every second on dev, so expired entries are evicted and reported on time. The new test defers a missing hash with a one second cutoff and checks that it is evicted and reported without anything else being added.
__________
Type: commit
Hash: e12f1c376f06704ff00809e745793d26540cf3be
Created: today
__________
gr0v1ty/nano-node:e12f1c376f06704ff00809e745793d26540cf3be
| Test Case | Status | Duration | Median Deviation | Log | Graph |
|---|---|---|---|---|---|
| 5n4pr_conf_10k_bintree | PASS | 95 s | -14.0 s | GitHub job logs | Show |
| 5n4pr_conf_10k_change | PASS | 112 s | -16.0 s | GitHub job logs | Show |
| 5n4pr_conf_change_dependant | PASS | 102 s | -12.0 s | GitHub job logs | - |
| 5n4pr_conf_change_independant | PASS | 102 s | -7.0 s | GitHub job logs | - |
| 5n4pr_conf_send_dependant | PASS | 97 s | -15.0 s | GitHub job logs | - |
| 5n4pr_conf_send_independant | PASS | 101 s | -13.0 s | GitHub job logs | - |
| 5n4pr_rocks_10k_bintree | PASS | 105 s | -6.0 s | GitHub job logs | - |
| 5n4pr_rocks_10k_change | PASS | 150 s | -6.0 s | GitHub job logs | - |