You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Oct 7, 2021. It is now read-only.
Currently, in ExecuteTestRunTask there is a sleep timer of about 3 seconds to give all other agents time to also receive this task. Provided there aren't too many problems, this helps to ensure that all agents start testing at roughly the same time (3 seconds being within the margin of error for most connections like iperf clients)
However this is obviously not going to scale. Need to re-introduce the concept of time sync for agents, and actually scheduling tests to take place, instead of relying on "execute" instructions from the server. Ideally, ExecuteTestRunTask could be removed, and InstallTestRunTask will be responsible for setting up a scheduled time for execution. This means that as soon as the agent receives it's InstallTestRunTask, it is autonomous until the test is completed.
This also lends itself to the whole "testing parallelization" question. You should be able to at least schedule multiple tests simultaneously, if not run them simultaneously. Need to think about that goal (as well as consolidate other issues that might have to do with this).
The text was updated successfully, but these errors were encountered:
Currently, in
ExecuteTestRunTask
there is a sleep timer of about 3 seconds to give all other agents time to also receive this task. Provided there aren't too many problems, this helps to ensure that all agents start testing at roughly the same time (3 seconds being within the margin of error for most connections like iperf clients)However this is obviously not going to scale. Need to re-introduce the concept of time sync for agents, and actually scheduling tests to take place, instead of relying on "execute" instructions from the server. Ideally,
ExecuteTestRunTask
could be removed, andInstallTestRunTask
will be responsible for setting up a scheduled time for execution. This means that as soon as the agent receives it'sInstallTestRunTask
, it is autonomous until the test is completed.This also lends itself to the whole "testing parallelization" question. You should be able to at least schedule multiple tests simultaneously, if not run them simultaneously. Need to think about that goal (as well as consolidate other issues that might have to do with this).
The text was updated successfully, but these errors were encountered: