BIND 10 #1342: Reintroduce delayed restarts after #213

[BIND 10 Development]

#1342: Reintroduce delayed restarts after #213
-------------------------------------+-------------------------------------
                   Reporter:         |                 Owner:  vorner
  vorner                             |                Status:  reviewing
                       Type:         |             Milestone:
  defect                             |  Sprint-20111122
                   Priority:  major  |            Resolution:
                  Component:  Boss   |             Sensitive:  0
  of BIND                            |           Sub-Project:  Core
                   Keywords:         |  Estimated Difficulty:  4
            Defect Severity:  N/A    |           Total Hours:  0
Feature Depending on Ticket:         |
        Add Hours to Ticket:         |
                  Internal?:  0      |
-------------------------------------+-------------------------------------
Changes (by jelte):

 * owner:  jelte => vorner


Comment:

 Replying to [comment:9 vorner]:
 > Hello
 >
 > Replying to [comment:7 jelte]:
 >
 > I fixed the boss tests to pass (and pushed), as you removed the
 dead_processes variable.
 >

 ah, doh, thanks, I had that change too, but had edited the .py, and not
 noticed the .in file :/

 > I guess you see no way to test the boss code itself as unittest? Anyway,
 you might want to check how components interacts with various kinds and
 being killed sooner or later.
 >

 There may be a way to do it, I may be able to come up with a mocky
 thingy, but either that would mostly be testing the mock itself, and I
 think having a test that actually kills a process from the outside
 would be more useful.

 > Actually, I found a problem with the behaviour. If I choose a needed
 component (let's say b10-auth in the default configuration) and kill it
 later than 10s after the system started, it is restarted. If I kill it
 again soon afterwards, the system stops. However, the semantics of needed
 component was to stop only if it fails at startup (with 10s being large
 enough time to fit the initialization in there probably), when the user
 typed the command and can see the system is ill. But this might happen
 after the user went for a holiday, so in this case it shouldn't bring the
 system down.
 >

 Heh. I *knew* it was too convenient that that variable existed already
 :)

 Anyway, I added a new one, _original_start_time which is now used for
 the check for initial failure (and which is only set on the very first
 call to start()). This should now be fixed.

 > Also, what is the purpose of the restart_delay parameter? Nothing uses
 it.
 >

 seemed useful, but yeah, if we don't use it we might as well hardcode
 the value. done.

 > And, the description of this parameter is misleading. It assumes the
 component is restarted every time, no matter if the component might be
 core for example.
 >

 specific comment isn't relevant anymore, as i just removed the
 parameter, but i added some new text, please take a look at it

-- 
Ticket URL: <http://bind10.isc.org/ticket/1342#comment:10>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development