[jira] Updated: (DERBY-2921) Replication: Add a network service that connects the master and slave Derby instances

["V.Narayanan (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/DERBY-2921?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

V.Narayanan updated DERBY-2921:
-------------------------------

    Attachment: Replication_Network_v4.stat
                Replication_Network_v4.diff

Thank you for the detailed responses Oystein,

I have addressed comments 4-19 and am working 
on the comments 1, 2 and 3. I will address 1, 2 and 3
and post a follow up again.

Some related responses

-19

* Removed the spaces between the names of the functions and the parenthesis. 
  The space between the parenthesis and the keywords however remain.

-18

* Will it be OK if I write unit test in a follow-up patch to this issue. This
  code is not being used for now and I feel it is harmless for now.

-17

* For now TYPE_ERROR messages are used only during the initial exchange while
  exchanging UID versions. This did not contain a SQLState to be transmitted.
  I think this is a good idea and I will use this when a SQLException is actually
  thrown.

-16

* I take this comment as to mean that the javadoc should explicitly define what
  kind of object needs to be transmitted as part of this ReplicationMessage. I
  will update the javadoc to indicate what type flag corresponds to what type
  of message. I shall also update the type flag definitions to define what type
  of message each of them signify

  For example

  public static final int TYPE_LOG = 0;

  will be modified to

  //This flag will be used for all messages will carry LogRecords.
  public static final int TYPE_LOG = 0;

  If any other action was expected of me here I sincerely apologize and request
  further elucidation of the action suggested.

- 15

* I have modified the class level comment to say the following

  This message is used for the communication between the master and the 
  slave during Replication.

  Please do mention if you think this is not good enough, I will change it
  again and post as per your suggestion.

- 14

* ReplicationMessageReceive javadoc: "receives the message from 
  the slave ...".  Is this class only to be used by the master?

  Although the class might serve a more generic purpose after the suggested
  factoring out of the generic network related code and the replication code
  in the current context the comment is correct since the intention was to use
  it only as a receiver of messages on the master.

- 13

* I planned to use the version check in ReplicationMessage#readExternal at a later
  stage to be able to handle a heterogenous collection of Derby version code.

  I did not want to do version checking here during initiation because I need to be
  able to inform the master that it suffers from an incompatibility with the slave. I
  do not want to abruptly fail the slave without informing the master and if I start
  to inform the master for every message received it would be too much communication 
  overload.

  I thought this initiator message would allow me to establish message incompatibilities at
  startup and also

- 12

* I have changed the method to private.

- 11

- 10

* I have changed ReplicationMessageReceive to accept a hostName. So this would take care of both 11 and 10.

- 9

* Since the underlying code will throw this exception do you want me to change this comment?

- 8

* Removed the local variable send.

- 7

* Modified the javadoc to mention both ObjectInputStream and ObjectOuputStream

- 5

* changed oos to objOuputStream and ois to objInputStream in both the receiver and the transmitter.

- 4

* Moved the declarations as suggested.


> Replication: Add a network service that connects the master and slave Derby instances
> -------------------------------------------------------------------------------------
>
>                 Key: DERBY-2921
>                 URL: https://issues.apache.org/jira/browse/DERBY-2921
>             Project: Derby
>          Issue Type: Sub-task
>          Components: Services
>    Affects Versions: 10.4.0.0
>            Reporter: Jørgen Løland
>            Assignee: V.Narayanan
>         Attachments: Replication_Network_v1.diff, Replication_Network_v1.stat, Replication_Network_v2.diff, Replication_Network_v2.stat, Replication_Network_v3.diff, Replication_Network_v3.stat, Replication_Network_v4.diff, Replication_Network_v4.stat
>
>
> A network connection is required between the master and slave Derby instances of a replicated database. The connection will be used to send many kinds of messages, including:
> * log records
> * the database (when replication is started)
> * master -> slave commands (like "stop replication")

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.