Currenlty, the default behavior is to download the single most recent file for a single machine or for each task or machine depending on the other arguments.
There are two options: --all to download all files and --last n to download only the most recent n files where n defaults to 1.
Our machines will only ever have a single log file associated with them now and a user providing a --task_id with no other options will likely want/expect to receive the most recent logs associated with the task. These "most recent logs" might be split across multiple machines.
So, the default and, I argue, only behavior should be to download all of the logs associated with the most specific option provided. In other words:
If --machine_id is provided, download the single log file associated with that machine.
If --task_id is provided, download the log file for every machine associated with that task.
If --job_id is provided, download all log files belonging to the machines associated with every task that is associated with that job.
Currenlty, the default behavior is to download the single most recent file for a single machine or for each task or machine depending on the other arguments. There are two options:
--all
to download all files and--last n
to download only the most recentn
files wheren
defaults to1
.Our machines will only ever have a single log file associated with them now and a user providing a
--task_id
with no other options will likely want/expect to receive the most recent logs associated with the task. These "most recent logs" might be split across multiple machines.So, the default and, I argue, only behavior should be to download all of the logs associated with the most specific option provided. In other words: If
--machine_id
is provided, download the single log file associated with that machine. If--task_id
is provided, download the log file for every machine associated with that task. If--job_id
is provided, download all log files belonging to the machines associated with every task that is associated with that job.AB#167800