We experienced test failure for test_wt2853_perf in the new memory sanitizer job.
[2019/11/01 10:50:33.009] GAP: 8 secs after 129 gets [2019/11/01 10:50:33.009] ......*.........*.........*.........*.........*.........*.........*.........*.........*.........*G.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.....G....*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.......G..*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*..G.......*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........G*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.G........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*.........*......... [2019/11/01 10:50:33.009] thread 0 did 793 joins (24 fails) [2019/11/01 10:50:33.009] ERROR: 24 failures when a single commit took more than 7 seconds. [2019/11/01 10:50:33.009] This may indicate a real problem or a particularly slow machine. [2019/11/01 10:50:33.009] Running test command: ./test_wt2853_perf [2019/11/01 10:50:33.009] test_wt2853_perf: FAILED: main/191: nfail == 0 [2019/11/01 10:50:33.009] process aborting
This test was timing sensitive and previously in Jenkins we bypass it for valgrind job, and avoid running it on slow machines.