#1478
  • Created on
    2011-12-10 16:57
  • Reported by
    rkuhn
  • Status
    Fixed
  • Priority
    High (2)
  • Assigned to
    drewhk
  • Milestone
    2.2-M3
  • Plan Level
    -
  • Due Date
    -

Attachments

system messages shall “eventually” be delivered

remote supervision means that if the Terminate() or ChildTerminated() messages are lost, something will get stuck. These messages must probably be retransmitted in that case.

Also: remove publication of dead system messages to deadLetters once we have that guarantee in place, since they do not belong into user event handlers.

[I didn’t find it in my heart to delete this gem, so I’ll just move it down a bit]
possible scheme:

- supervisor sends Terminate()
- supervisor does not get ChildTerminated() after X time => reparent the child to DavyJones
- DavyJones periodically sends Terminate() and subscribes to DeathWatch for all its “children” and releases them once he gets any sort of reply
- supervisor forwards all ChildTerminated() he gets for “unknown” children to DavyJones
User picture
  on 2011-12-10 16:58 *
By rkuhn
we could easily translate “100 years” into “100 ms” here
User picture
  on 2011-12-10 20:09 *
By rkuhn
Description changed from remote supervision means th... to remote supervision means th...
User picture
  on 2011-12-27 17:40 *
By viktorklang
Milestone changed from 2.0-M2 to 2.0-M3
Updating tickets (#1190, #1322, #1375, #1406, #1408, #950, #1162, #1166, #1236, #1255, #1310, #1423, #1428, #1468, #1476, #1477, #1478, #1509, #1511, #1519, #1534, #1537, #1538, #1542, #1549, #1557, #1037, #1085, #1173, #1231, #1364, #1485, #1525, #1526, #1527, #1528, #1553, #1348, #1403)
User picture
  on 2012-01-17 20:45 *
By viktorklang
Milestone changed from 2.0-M3 to 2.0
Updating tickets (#1322, #1408, #1478, #1519, #1595, #1639, #1173, #1231, #1364, #1542, #1553, #1559, #1348)
User picture
  on 2012-02-23 15:08 *
By rkuhn
Milestone changed from 2.0 to 2.1
User picture
  on 2012-05-23 15:26 *
By Jonas Bonér
Milestone changed from Coltrane to Mingus
Updating tickets (#644, #1478, #1596, #1715, #1978, #1988, #849)
User picture
  on 2012-07-05 22:52 *
By viktorklang
Priority changed from Normal (3) to Low (4)
User picture
  on 2012-09-24 14:41 *
By viktorklang
Milestone changed from 2.1 to Coltrane
User picture
  on 2012-11-01 18:04 *
By viktorklang
Priority changed from Low (4) to High (2)
User picture
  on 2012-11-09 14:48 *
By rkuhn
Status changed from New to Accepted
User picture
  on 2012-11-09 14:48 *
By rkuhn
Assigned to set to rkuhn
User picture
  on 2012-11-09 20:32 *
By rkuhn
a solution which looks promising needs the new remoting:

  • in the writer actor for each connection, wrap SystemMessage/Failed/Terminated inside sequenced envelopes which are retried/deduplicated
  • unless I’m currently missing something, this should not observably violate ordering guarantees
  • no holding of traffic necessary at any point, i.e. only the retry timer
User picture
  on 2012-12-18 17:05 *
By rkuhn
Description changed from remote supervision means th... to remote supervision means th...
User picture
  on 2012-12-20 14:55 *
By rkuhn
Milestone changed from Current to Coltrane
User picture
  on 2013-01-21 18:43 *
By rkuhn
Estimate changed from None to Large
Sum of child estimates changed from 0.0 to 7.0
User picture
  on 2013-03-27 16:21 *
By drewhk
Assigned to changed from rkuhn to drewhk
User picture
  on 2013-03-27 16:21 *
By drewhk
Milestone changed from Coltrane to Current
User picture
  on 2013-03-27 22:15 *
By drewhk
So I implemented now a retransmission scheme (a pair of smart buffers) with sliding window and selective NACKs for efficiency. I stress-tested it, too.

These are only the building blocks, it is not hooked yet into the remoting. Also, it is currently somewhat ugly. Proceeding with cleanup and then integration with remoting.
User picture
  on 2013-04-15 21:12 *
By drewhk
Status changed from Accepted to Test
User picture
  on 2013-04-19 01:08 *
By drewhk
Status changed from Test to Fixed