git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@809 71dcc9de-d417-0410-9af5-da40c76e7ee4

Make client asynchronous base handler only attemp to dispatch response packets, and raise if it is not an expected response.
Add a method on Packet class to determine if it's a response.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@808 71dcc9de-d417-0410-9af5-da40c76e7ee4

Make MTClientConnection.ask register message to dispatcher. This makes sure registration happens before the message has a chance of being sent.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@807 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@806 71dcc9de-d417-0410-9af5-da40c76e7ee4

replication.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@805 71dcc9de-d417-0410-9af5-da40c76e7ee4

nodes to other storage nodes (during replications). Move The handler method in
the common handler class. Move AskTransactionInformation handler in the same
handler for the same reason and clean the same method in the verification
handler, where the message can only be sent by the primary master node. 
Fix a super construtor call.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@804 71dcc9de-d417-0410-9af5-da40c76e7ee4

Prevent killing the ast connection iterated upon when no connection has been closed (kills primary master node connection when any storage is in a non-running state upon startup...).


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@803 71dcc9de-d417-0410-9af5-da40c76e7ee4

decorators. Delete methods that raise UnexpectedPacketError since the default
base handler already do this job.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@802 71dcc9de-d417-0410-9af5-da40c76e7ee4

way to clean old nodes on non-master nodes.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@801 71dcc9de-d417-0410-9af5-da40c76e7ee4

the setClusterState packet with is received from the admin node. If no partition
table/IDs were gathered before, wait for at least one storage node and use the
first to create the initial partition table.
A timeout is left to act as it was done before, to preserve automatic startup in
tests, will be removed latter.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@800 71dcc9de-d417-0410-9af5-da40c76e7ee4

Make handleAnswerPrimaryMaster, handleRequestNodeIdentification and handleAskPrimaryMaster resilient to connection loss. As they are involved in question-response exchanges during election on connections which will get dropped as soon as a master arises, they must check if the connection is still opened before trying to send.
Note: the connection might get closed anyway during the actual send, but this is handled separately already. The problem we fix here emerged because of more tedious checkings in Connection's ask & answer methods.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@799 71dcc9de-d417-0410-9af5-da40c76e7ee4

consistent with client <> master communication.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@798 71dcc9de-d417-0410-9af5-da40c76e7ee4

Don't let polling thread die from any exception. Instead, emit a log with exception traceback and retry polling.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@797 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@796 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@795 71dcc9de-d417-0410-9af5-da40c76e7ee4

The recovery stage is used only one time, when switching to primary state.
Each time the cluster lost the operational status (a cell has no up to date
node), the verification stage restart.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@794 71dcc9de-d417-0410-9af5-da40c76e7ee4

storage is gathering the full table. 


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@793 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@792 71dcc9de-d417-0410-9af5-da40c76e7ee4

already known by UUID. This avoid duplicate entries in the node manager.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@791 71dcc9de-d417-0410-9af5-da40c76e7ee4

broadcast it as down state. Just broadcast its information and running state.
Master don't send the partition table to storage node, they are it from the
initialization storage stage.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@790 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@789 71dcc9de-d417-0410-9af5-da40c76e7ee4

Move common handlers in handler.py module.
Add InitializationHandler, applied on the master connection while retreiving
the partition table and node list.
Replace a call to (inexistant) method handleUnexpectedPacket with a raise of
UnexpectedPacketError exception.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@788 71dcc9de-d417-0410-9af5-da40c76e7ee4

- test more steps
- fix erroneous assertion
- use a local variable to detect that all handler were executed, rather than a class variable


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@787 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@786 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@785 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@784 71dcc9de-d417-0410-9af5-da40c76e7ee4

handlers/replication.py in commit #778.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@783 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@782 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@781 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@780 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@779 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@778 71dcc9de-d417-0410-9af5-da40c76e7ee4

directory. Replicator's handler is moved into handlers/replication.py. 
Clean some imports.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@776 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@775 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@774 71dcc9de-d417-0410-9af5-da40c76e7ee4

- Split completely expected packet handlers from asynchronous event handlers
- Remove NEOStorageConnectionFailure exception and replace it by a more general "ConnectionClosed" exception, purely internal to client app.
- Modify waitMessage to redirect to a handler based on remote peer type
- Fix multiple code paths taken when master doesn't accept connections (or closes them) while trying to conenct to it.
Also:
- remove unused imports found by pyflakes


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@773 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@772 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@771 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@770 71dcc9de-d417-0410-9af5-da40c76e7ee4

Add lock/release around each "close" call in client which wasn't already locked.


git-svn-id: https://svn.erp5.org/repos/neo/branches/prototype3@769 71dcc9de-d417-0410-9af5-da40c76e7ee4