The Evergreen linux-directio task failed periodically on the waterfall page, hitting the default 2 hours timing out for the Evergreen shell. Some works had been done in WT-5255 and WT-5450 providing a new major_timeout option in the format.sh script which could be helpful.
[2020/01/18 04:04:57.981] for i in $(seq 3); do [2020/01/18 04:04:57.981] ./t -1 -c ../../../test/format/CONFIG.stress -C "direct_io=[data]" 2>&1 [2020/01/18 04:17:35.126] t: process 23783 running [2020/01/18 04:17:35.126] 1: table, row-store (758 seconds) [2020/01/18 04:17:35.126] t: successful run completed [2020/01/18 04:22:33.554] t: process 24012 running [2020/01/18 04:22:33.554] 1: table, row-store (298 seconds) [2020/01/18 04:22:33.554] t: successful run completed [2020/01/18 06:22:33.807] Command stopped early: context canceled [2020/01/18 06:22:33.807] Running task-timeout commands.