If I got you right, then what you want is something that has the exact lifetime of the agent process and allows you any kind of a custom API.
Maybe you should consider using a module for this. There is this notion of so called anonymous modules. Such module simply coexists with the parent process, in your case the agent. See http://mesos.apache.org/documentation/latest/modules/ <http://mesos.apache.org/documentation/latest/modules/> for more. Such module in combination with a custom executor might get you where you need to be…. at least that is what I am hoping for. > On Aug 9, 2017, at 12:53 PM, Oeg Bizz <[email protected]> wrote: > > Ben, > That is exactly what I did, the problem is that is not being invoked when > the mesos-slave starts. I am trying to avoid sending a bogus initial task > just to kick it off, but it seems there is no way to tell the mesos-slave to > start the executor at startup, is it? > > Thanks for your reply, > > Oscar > > > On Tuesday, August 8, 2017, 3:38:17 PM EDT, Benjamin Mahler > <[email protected]> wrote: > > > You're free to write your own long lived executor that can process multiple > tasks. The built in executors self-terminate after running the tasks they are > launched with. > > On Tue, Aug 8, 2017 at 2:36 AM, Oeg Bizz <[email protected] > <mailto:[email protected]>> wrote: > It is used to notify some services that the agents are out there ready to > process data. The service takes into account who and how many are out there > and some information from the HB to perform some sort of load balancing. > Without the HB the service does not know how many agents are out and tasking > is assigned only when an offer is received, but that is not as effective as > it should be if the service would know in advance about the cluster's load. > Thanks, > > Oscar > > > On Monday, August 7, 2017, 11:13:10 AM EDT, James Peach <[email protected] > <mailto:[email protected]>> wrote: > > > > > On Aug 5, 2017, at 3:03 AM, Oeg Bizz <[email protected] > > <mailto:[email protected]>> wrote: > > > > I have a framework that relies on information sent by a custom Java Command > > Executor; think of some sort of heartbeat. I start getting hearbeats after > > I send a task to that mesos-slave, but never before that. That makes me > > assume that the CommandExecutor is not started until a task is submitted to > > be executed by that agent. Is there a way to tell mesos-slave to start the > > ComandExecutor as soon as it starts running? > > Not AFAIK. Executors are always spawned in order to execute tasks. In your > case, what is the heartbeat for, if there are no tasks on the agent? > > > J >
