Right now, we write our cot log to public/logs/chain_of_trust.log. If this passes, then we start writing public/logs/live_backing.log using the task output. This works in the standard case, where cot verification passes.
When cot fails, treeherder doesn't show the cot failure, because it doesn't know about chain_of_trust.log. We could solve this by teaching treeherder about this log, or we could solve by:
-
writing the chain of trust log to public/logs/live_backing.log, and EITHER
-
moving the chain of trust log to public/logs/chain_of_trust.log on success, OR
-
appending the task log to the chain of trust log at public/logs/live_backing.log, depending on whether we prefer 2 different log files or not.
Right now, we write our cot log to
public/logs/chain_of_trust.log. If this passes, then we start writingpublic/logs/live_backing.logusing the task output. This works in the standard case, where cot verification passes.When cot fails, treeherder doesn't show the cot failure, because it doesn't know about
chain_of_trust.log. We could solve this by teaching treeherder about this log, or we could solve by:writing the chain of trust log to
public/logs/live_backing.log, and EITHERmoving the chain of trust log to
public/logs/chain_of_trust.logon success, ORappending the task log to the chain of trust log at
public/logs/live_backing.log, depending on whether we prefer 2 different log files or not.