natas_hw
/
cassandra
mirror of https://github.com/apache/cassandra
0
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

switch to Map<Long, List> to avoid re-hashing the list with every addition 

patch by jbellis; reviewed by thobbs for CASSANDRA-4287
This commit is contained in:
Jonathan Ellis 2012-05-25 15:16:52 -05:00
parent 6ab02c5997
commit 271761038c
2 changed files with 15 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
CHANGES.txt
View File

@ -3,6 +3,8 @@
* ensure unique streaming session id's (CASSANDRA-4223)
* kick off background compaction when min/max thresholds change
(CASSANDRA-4279)
* improve ability of STCS.getBuckets to deal with 100s of 1000s of
sstables, such as when convertinb back from LCS (CASSANDRA-4287)
1.0.10

26
src/java/org/apache/cassandra/db/compaction/SizeTieredCompactionStrategy.java
View File

@ -122,7 +122,7 @@ public class SizeTieredCompactionStrategy extends AbstractCompactionStrategy
}
});
Map<List<T>, Long> buckets = new HashMap<List<T>, Long>();
Map<Long, List<T>> buckets = new HashMap<Long, List<T>>();
for (Pair<T, Long> pair: sortedFiles)
{
@ -132,19 +132,19 @@ public class SizeTieredCompactionStrategy extends AbstractCompactionStrategy
// look for a bucket containing similar-sized files:
// group in the same bucket if it's w/in 50% of the average for this bucket,
// or this file and the bucket are all considered "small" (less than `minSSTableSize`)
for (Entry<List<T>, Long> entry : buckets.entrySet())
for (Entry<Long, List<T>> entry : buckets.entrySet())
{
List<T> bucket = entry.getKey();
long averageSize = entry.getValue();
if ((size > (averageSize / 2) && size < (3 * averageSize) / 2)
|| (size < minSSTableSize && averageSize < minSSTableSize))
List<T> bucket = entry.getValue();
long oldAverageSize = entry.getKey();
if ((size > (oldAverageSize / 2) && size < (3 * oldAverageSize) / 2)
|| (size < minSSTableSize && oldAverageSize < minSSTableSize))
{
// remove and re-add because adding changes the hash
buckets.remove(bucket);
long totalSize = bucket.size() * averageSize;
averageSize = (totalSize + size) / (bucket.size() + 1);
// remove and re-add under new new average size
buckets.remove(oldAverageSize);
long totalSize = bucket.size() * oldAverageSize;
long newAverageSize = (totalSize + size) / (bucket.size() + 1);
bucket.add(pair.left);
buckets.put(bucket, averageSize);
buckets.put(newAverageSize, bucket);
bFound = true;
break;
}
@ -154,11 +154,11 @@ public class SizeTieredCompactionStrategy extends AbstractCompactionStrategy
{
ArrayList<T> bucket = new ArrayList<T>();
bucket.add(pair.left);
buckets.put(bucket, size);
buckets.put(size, bucket);
}
}
return new ArrayList<List<T>>(buckets.keySet());
return new ArrayList<List<T>>(buckets.values());
}
private void updateEstimatedCompactionsByTasks(List<AbstractCompactionTask> tasks)