Sometimes a task (typically 2008 papers from UofA) may run longer than 10 minutes. This causes the tasks to be killed and retried later. The current version always tries the same task again on the next invocation, which is not good, because other tasks do not get scheduled if the problem continues.
Sometimes a task (typically 2008 papers from UofA) may run longer than 10 minutes. This causes the tasks to be killed and retried later. The current version always tries the same task again on the next invocation, which is not good, because other tasks do not get scheduled if the problem continues.