com.bigdata.sparse.TimestampChooser Maven / Gradle / Ivy
/*
Copyright (C) SYSTAP, LLC DBA Blazegraph 2006-2016. All rights reserved.
Contact:
SYSTAP, LLC DBA Blazegraph
2501 Calvert ST NW #106
Washington, DC 20008
[email protected]
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; version 2 of the License.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*/
/*
* Created on Aug 4, 2008
*/
package com.bigdata.sparse;
import com.bigdata.btree.BTree;
import com.bigdata.btree.IIndex;
import com.bigdata.btree.ILocalBTreeView;
import com.bigdata.journal.IJournal;
import com.bigdata.journal.ITimestampService;
import com.bigdata.journal.TemporaryRawStore;
import com.bigdata.sparse.TPS.TPV;
import com.bigdata.util.MillisecondTimestampFactory;
/**
* Utility class for choosing timestamps for the {@link SparseRowStore} on the
* server.
*
* @author Bryan Thompson
* @version $Id$
*/
public class TimestampChooser implements IRowStoreConstants {
/**
* Choose the timestamp for {@link TPV} tuples to written on the sparse row
* store.
*
* Note: Revisions written with the same timestamp as a pre-existing column
* value will overwrite the existing column value rather than causing new
* revisions with their own distinct timestamp to be written. There is
* therefore a choice for "auto" vs "auto-unique" for timestamps.
*
* Note: Timestamps generated locally on the server will be consistent
* within a row, and all revisions of column values for the same row will
* always be in the same index partition and hence on the same server. This
* means that we can use locally assigned timestamp as unique timestamps.
* However, time could go backwards if there is a failover to another server
* for the partition and the other server has a different clock time. This
* is resolved by choosing a timestamp assigned by the global
* {@link ITimestampService}.
*
* @param timestamp
* When timestamp is {@link IRowStoreConstants#AUTO_TIMESTAMP}
* the timestamp is the local system time. When timestamp
* is {@link IRowStoreConstants#AUTO_TIMESTAMP_UNIQUE} a federation
* wide unique timestamp is assigned by the
* {@link ITimestampService}. Otherwise the caller's
* timestamp is returned.
*/
static public long chooseTimestamp(final IIndex ndx, final long timestamp) {
if (timestamp == AUTO_TIMESTAMP) {
return System.currentTimeMillis();
} else if (timestamp == AUTO_TIMESTAMP_UNIQUE) {
/*
* The BTree that is absorbing writes.
*/
final BTree mutableBTree = ((ILocalBTreeView) ndx)
.getMutableBTree();
if(mutableBTree.getStore() instanceof TemporaryRawStore) {
/*
* Use a unique timestamp for the local machine since a
* temporary store is not visible outside of that context.
*/
return MillisecondTimestampFactory.nextMillis();
}
/*
* The backing store will be some kind of IJournal - a Journal, a
* ManagedJournal, an IsolatedActionJournal, etc.
*/
final IJournal journal = (IJournal) mutableBTree.getStore();
/*
* This will be locally unique for a Journal and federation-wide
* unique for a ManagedJournal. In the former case the timestamp is
* assigned locally. In the latter case it will use a robust method
* to discover the timestamp service and obtain a federation-wide
* unique timestamp.
*/
// try {
return journal.getLocalTransactionManager().nextTimestamp();
// } catch(IOException ex) {
//
// /*
// * Note: Declared for RMI interoperability.
// */
//
// throw new RuntimeException(ex);
//
// }
} else {
// return the caller's value.
return timestamp;
}
}
}