BIND 10 #1847: CC timeouts in auth

[BIND 10 Development]

#1847: CC timeouts in auth
-------------------------------------+-------------------------------------
                   Reporter:  jreed  |                 Owner:
                       Type:         |                Status:  closed
  defect                             |             Milestone:  New Tasks
                   Priority:         |            Resolution:  invalid
  medium                             |             Sensitive:  0
                  Component:         |           Sub-Project:  DNS
  Unclassified                       |  Estimated Difficulty:  2
                   Keywords:         |           Total Hours:  0
            Defect Severity:  N/A    |
Feature Depending on Ticket:         |
        Add Hours to Ticket:  0      |
                  Internal?:  0      |
-------------------------------------+-------------------------------------

Comment (by jinmei):

 Replying to [comment:10 vorner]:

 > > I'd actually suggest revisiting the IPC method for statistics at more
 > > fundamental level.  Blocking/timeout issue aside, I suspect the CC
 > > channel is not appropriate for exchanging statistics information due
 > > to scalability issues.  I assume we'd eventually (soon?) like to be
 > > able to have per-zone statistics, while still allowing millions of
 > > zones to be served.  If there's a reasonable way of using the CC
 > > channel to exchange all of the statistics, that's fine, but I suspect
 > > we need to use a different method for this purpose.
 >
 > I'm not sure what you mean. If it's some way to push only diffs of the
 statistics, I think we want that sometime ‒ if we have millions of zones,
 its unlikely all of them would be accessed every 10 seconds or like that,
 so it makes little sense to send them all. Having a poll principle instead
 of push or notification or something would seem better too.
 >
 > But if you mean creating YAUDS (Yet Another Unix Domain Socket) for the
 purpose of statistics, I'd be strongly against that. I believe the CC
 channel should be checked and made sure it can handle many and large
 messages to cope with the load. Having millions of zones would create
 messages several megabytes large, but these should pose no problem for the
 system.

 I was not proposing anything specific at this moment, and in any event
 it wouldn't be YAUDS (depending on what exactly and specifically you
 mean YAUDS).  One possibility I had in my mind is that b10-auth stores
 statistics in a separate storage such as an SQLite3 DB file, maybe in
 a periodic batch task.  b10-stats or any other statistics clients
 would then retrieve the stored statistics values from that storage
 separately.

 I'm not necessarily excluding the use of the common message bus,
 however, as long as we can exchange necessary information in a
 reasonably timely fashion without disrupting the main server of
 b10-auth (i.e., responding to queries).

-- 
Ticket URL: <http://bind10.isc.org/ticket/1847#comment:12>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development