When member joins the group and needs to receive an SST it won't
receive the corresponding menbership view event because the SST
happens after the event and will already include the effects of
all events ordered before it. The view then must be recovered from
the received state.
Minor renames and cleanups.
References codership/wsrep-lib#18
Dbsim has internal map of server objects for SST simulation.
This was mapped using server_id, which is not available
anymore when server object is constructed. Changed the dbsim to
use server name instead for internal mapping.
it on disconnect.
- Don't rely on own index from the view because the view may come from
another member (IST/SST), instead always determine own index from own ID.
Refs codership/wsrep-lib#13
with superproject definitions.
Avoid using client_service::do_2pc() in before_commit() to
determine if 2pc is actually happening, will use transaction
states to deduce that. client_service::do_2pc() should be deprecated.
Fixed a compiler warning in db_high_priority_service.cpp.
* Check fragment removal error code in prepare phase. It is possible
that the transaction gets BF aborted during fragment removal.
* Mark fragment certified in certify_fragment() even if the provider
returns cert failed error. With current wsrep-API error codes
it may not be possible to distinquish certification failure
and BF abort during fragment replication. This may also be a
provider bug. As a result rollback fragment may sometimes be
replicated when it would not be necessary.
* Enable codepath to BF abort high priority SR applier
* Pass ws_handle, ws_meta to high priority service rollback
call to allow total ordering of rollback process
The write set handle and meta data are needed for SR transactions
where the commit context is not known when the transaction starts.
The passed handle and meta data can be set through client_state
prepare_for_ordering() call before performing commit.
The intended purpose for local mode was local storage access without
entering replication hooks. However, the same can be achieved with
high priority mode. Removed replicating mode and use local instead to
denote locally processing clients.
quite useful as there might not be enough information for it
after the statement has been processed. Better to handle retrying
on DBMS side. Also removed after_statement_result enumeration and
return plain int from after_statement().
The interface method can be used to notify the DBMS implementation
about state changes in well defined order. The call will be done
under server_state mutex protection.
* Pass condition variable for client_state
* Notify all cond waiters when changing the transcation status to
aborted
* Wait for aborting transaction state aborted in before_command
* Added bootstrap service call to do DBMS side bootstrap operations
during the cluster bootstrap.
* Added last_committed_gtid() to provider interface
* Implemented wait_for_gtid() provider call
* Pass initial position to the server state
* Propagate server max protocol version to provider init options
* Store gtid from connected call to make cluster id and the connect
position available
Depending on the DBMS client session allocation strategy the
client id may or may not be available when the client_session
is constructed, therefore there should be a method to assign
an id after construction. Close/cleanup methods were added to
clean up open transactions appropriately.
