Re: [lmap] A question about running actions in sequence

[Juergen Schoenwaelder <[e-mail]>]

On Tue, May 16, 2017 at 05:35:36PM +0200, Fredrik Kers wrote:
> Hi!
> 
> I have a question about how running Actions in sequence is supposed to work
> according to the LAMP information model.
> 
> Let's take a simple example. Say you have two actions that you want to run
> in sequence, first a traceroute that finds all routers between the MA and a
> remote peer, and after that a time limited ping to each of the routers.
> 
> It will not be possible to use the Schedule's duration to do the time
> limiting of the ping since there is no way to have a second Schedule (doing
> the ping) starting after the first Schedule (doing the traceroute) is
> finished.

Right, the schedule's duration is a limit on the duration of the whole
schedule, which may consist of multiple actions.

> Also, you can not place both Actions in the same Schedule and use the
> Schedule's duration since you don't know how long time the traceroute will
> take.

Yes.

> The only way that I see is to make it possible to configure the duration in
> the ping Actions itself. Is this how it's intended to be done?

It seems this is what you are looking for. Right now, the information
model does not support per action duration limits. This could be added
though, I think, without breaking anything.

> The possibility to schedule arbitrary actions in a sequence would be
> quite limited by this I think.

I am not sure this statement is right. But yes, if you want to have
per action duration limits, they are not part of this version of the
LMAP models. Whether this is a major omission or not - well I do not
know.

Perhaps a somewhat kludgy workaround could be a schedule that does
just the traceroute (with a duration set to cover the traceroutes)
where the traceroute action feeds into another schedule (with a
duration set to cover pings) that does the pings.

/js

PS: Personally, I find the concrete use case a bit surprising. If you
    want to probe the path for a certain time, why not run traceroute
    for a certain time; this causes roughly the same number of probe
    packets and your measurement adjusts to any path changes during
    the measurement itself. But perhaps you just picked this as an
    example to illustrate the question.

-- 
Juergen Schoenwaelder           Jacobs University Bremen gGmbH
Phone: [phone]         Campus Ring 1 | 28759 Bremen | Germany
Fax:   [phone]         <http://www.jacobs-university.de/>