BIND 10 #930: update stats modules to get statistics list from cfgmgr, and define the list in spec file of each module
BIND 10 Development
do-not-reply at isc.org
Mon Aug 8 08:24:14 UTC 2011
#930: update stats modules to get statistics list from cfgmgr, and define the
list in spec file of each module
-------------------------------------+-------------------------------------
Reporter: | Owner: naokikambe
naokikambe | Status: reviewing
Type: | Milestone:
enhancement | Sprint-20110816
Priority: major | Resolution:
Component: | Sensitive: 0
statistics | Sub-Project: DNS
Keywords: | Estimated Difficulty: 20.0
Defect Severity: N/A | Total Hours: 0
Feature Depending on Ticket: |
Add Hours to Ticket: 0 |
Internal?: 0 |
-------------------------------------+-------------------------------------
Changes (by vorner):
* owner: vorner => naokikambe
Comment:
Hello
Replying to [comment:12 naokikambe]:
> > One is, if it is checked by the module which is sending the data, it
does not solve much. If it is buggy, it can as well do the validation
wrong and send wrong data anyway. And if the module wants to be malicious,
nothing stops it from being so. Therefore I don't think validation on the
sending side solves any problem (or, I at last don't see any problem being
solved by this).
>
> I'm very sorry if I am misunderstanding. Of course the stats module
validates statistics data received from other modules. Here the validation
is just a spec-format validation before sending statistics data. The
validation is done by a non-stats module itself (like auth, boss etc). For
example the implementation of this validation is like
ModuleSpec::validateConfig or like ModuleSpec::validateCommand already
implemented. And Jelte-san commented in #928 that statistics validation
also would be required for other modules.([comment:6:ticket:928])
Ah, I see. OK, then it makes sense.
> However actually more wait is required after _started.set() is done (in
other word after _started.wait() is done). As one solution, should we move
_started.set() to inside of the target module, for example, just before
the while-loop in the target module is started? But it might be not good
to customize inside of the target module for test only.
OK, let's leave it this way for now, because it doesn't look trivial. But
could you, please, create a ticket to get rid of that sleep somehow? I
still think it is not really nice to have it there, though I don't see an
easy way to avoid it right now.
So, I believe the branch is ready.
--
Ticket URL: <http://bind10.isc.org/ticket/930#comment:13>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development
More information about the bind10-tickets
mailing list