BIND 10 #1514: Update SERIAL in DDNS

[BIND 10 Development]

#1514: Update SERIAL in DDNS
-------------------------------------+-------------------------------------
                   Reporter:  jelte  |                 Owner:  haikuo
                       Type:  task   |                Status:  reviewing
                   Priority:         |             Milestone:
  medium                             |  Sprint-20120612
                  Component:  DDNS   |            Resolution:
                   Keywords:         |             Sensitive:  0
            Defect Severity:  N/A    |           Sub-Project:  DNS
Feature Depending on Ticket:  DDNS   |  Estimated Difficulty:  4
        Add Hours to Ticket:  0      |           Total Hours:  0
                  Internal?:  0      |
-------------------------------------+-------------------------------------

Comment (by haikuo):

 Replying to [comment:12 jelte]:
 > Replying to [comment:11 haikuo]:
 >
 > The changes done look ok,
 >
 > >
 > > sorry, jelte, I maybe disagree about this suggestion. :)
 >
 > That is allowed :)
 >
 > > The rfc 2136 said the update operation should be ignore if the serial
 number is smaller than the old one. If the new serial number is smaller
 than the old one, the serial number will be increased according your codes
 above.
 >
 > Unless you are looking at a different section than I am, or I am
 misreading it, the RFC says to ignore the specific SOA update (not the
 entire update). It also says in a different place that if the SOA is not
 updated in the entire update, it should be autoincremented.
 >
 > To put it another way, the way I read it is that the SOA must be auto-
 incremented, except if the update contains a SOA RR with a higher serial.
 >
 > If the serial is lower, it is ignored, and if it is ignored, the SOA
 would not be updated at all, in which case it would be a violation of the
 'must autoincrement if not done by update'.
 >
 > And even if that is not true, the current code would result in SERVFAIL
 if you give it a lower serial (since it will try to use new_soa which has
 not been set to anything) :)
 >
 jelte, the RFC 2136 section 3.6 said each update operation should increase
 the soa number automatically, but it does't say the soa number should
 increase if the operation failed.
 I tested it using bind9 for this situation above, it looks like that the
 SOA number was not increased when the update operation failed.
 so  I need your feedback on how we do in this situation.

-- 
Ticket URL: <http://bind10.isc.org/ticket/1514#comment:13>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development