============================================
#opendaylight-clustering: “streaming design”
============================================


Meeting started by moizer at 17:18:35 UTC.  The full logs are available
at
http://meetings.opendaylight.org/opendaylight-clustering/2014/_streaming_design_/opendaylight-clustering-_streaming_design_.2014-10-28-17.18.log.html
.



Meeting summary
---------------

* all protocol buffer messages that send across a normalized node should
  now be migrated to a java serializable which implements custom
  readObject/writeObject  (moizer, 17:21:09)
* Saving snapshot in the replicated log uses up too much memory
  (moizer, 17:30:32)
* maybe we need to create a snapshot when we need to Install the
  snapshot  (moizer, 17:30:54)
* when shard is not persistent - there is more chance of install
  snapshot being required  (moizer, 17:31:46)
* AGREED: We will capture shot when we need to install it  (moizer,
  17:44:00)
* we want to keep serialization of instance identifiers on
  WriteData/MergeData with the new streaming  (moizer, 17:47:43)
* create a bug for instance identifier serialization  (moizer, 17:48:48)
* ACTION: create a bug for snapshotting/install snapshot to use the new
  streaming classes  (moizer, 17:50:33)
* Parallel Task 1 - Install snapshot changes with ReadDataReply
  (moizer, 17:52:24)
* Parallel Task 2 - WriteData/MergeData - change streamwriter to write
  instance identifiers  (moizer, 17:53:09)
* Parallel Task 3 - Add more unit tests for the stream writer  (moizer,
  17:53:42)
* Task 4 - Optimize the stream writer to reuse the builder - definitely
  the leafnode/leafsetentry builders can be reused  (moizer, 17:54:55)
* Task 5 - ReplicatedLogEntry also needs to use a java serializable
  (moizer, 17:56:28)
* We need to take care of versioning in the java serializable  (moizer,
  17:57:08)
* ACTION: evaluate the work required to support backward compatibility
  (moizer, 18:03:59)
* by backward compatibility I mean each message should be backward
  compatibility  (moizer, 18:05:43)
* What should we use as the version?  (moizer, 18:09:03)
* A static field in the clustering-messages could be enough  (moizer,
  18:09:23)
* 1. NormalizedNode , instead of bytestring , will be stored as part of
  the snapshot.  (moizer, 18:10:21)
* 2. Installsnapshot will use the NormalizedStreamWrite  (moizer,
  18:10:35)
* 3. All messages will have a version number. An actor only accepts
  messages with version >= its version.  (moizer, 18:10:46)
* 4. Its possible for a leader to have an higher version than the
  follower. In a 3 member cluster, when the first member is upgraded,
  its not possible for the leader > follower. But when the second member
  is upraded, it can happen that the one of the upgraded ones become the
  leader.  (moizer, 18:10:57)
* 5. Backward compatibility is supported. If a V1.1 node gets V1.0
  message , it should honor it.  (moizer, 18:11:05)
* 6. If a V1.0 node gets V1.1 message, it would ignore the message. The
  V1.0 replica, if follower, will be behind the Leader, till upgraded.
  The V1.0 replica , if Leader, will not honor messages from upgraded
  V1.1 follower.  (moizer, 18:11:13)
* 7. Heartbeats should be counted , even if messages are not honored.
  (moizer, 18:11:27)
* 8. VoteRequest should also comply (version >= its version) rule, there
  by ensuring that the leader is elected from the majority with same
  version  (moizer, 18:11:38)
* These are rules which Kamal proposes  (moizer, 18:11:53)
* When we get a message with an older version that can be handled the
  reply needs to be of the same version as the request  (moizer,
  18:15:33)
* all fields in our java serializable should be transient  (moizer,
  18:23:37)
* we may need to create message classes to cleanly migrate from v1 -> v2
  messages  (moizer, 18:26:57)
* Send message here asking about backward compatibility here ->
  https://lists.opendaylight.org/mailman/listinfo/opendaylight-users
  (moizer, 18:33:22)
* Message to users  (moizer, 18:36:14)
* We are changing the serialization of clustering messages to improve
  performance for persistence and communication.  (moizer, 18:37:34)
* Persisted data will also not be recovered due to this incompatibility
  (moizer, 18:39:25)
* Is this acceptable?  (moizer, 18:39:36)
* Title : MD-SAL Clustering for Helium  (moizer, 18:41:42)
* MD-SAL Clustering for Helium R1  (moizer, 18:41:54)
* Final title : MD-SAL Clustering changes for Helium R1  (moizer,
  18:42:16)
* Response requested by COB Friday  (moizer, 18:42:31)
* ACTION: Moiz to send the message  (moizer, 18:42:56)
* We will proceed with the assumption that backward compatibility is not
  to be supported  (moizer, 18:43:30)



Meeting ended at 19:19:18 UTC.



People present (lines said)
---------------------------

* moizer (44)
* odl_meetbot (3)



Generated by `MeetBot`_ 0.1.4