Passed all regression tests and PerformanceEvaluation running with multiple region servers. table.jsp now displays the correct information
HRegionInfo:
- removed getTableNameFromRegionName and parseMetaRegionRow we have the information in the meta table, just use it.
HServerInfo:
- I had originally made some changes here but removed them. The only remaining changes are javadoc
MetaScanner:
- build region name using ZEROES instead of NINES. When you scan you need a row name that sorts before the first row rather than after the last row.
- scan using COLUMN_FAMILY_ARRAY instead of COL_REGIONINFO_ARRAY. This way you also get the server name and start code
- change api for MetaScannerVisitor so that processRow only gets the RowResult. If you have the RowResult you have everything you need.
HConnectionManager:
- change listTables' MetaScannerVisitor to conform to new processRow api
HTable:
- change getStartKeys' MetaScannerVisitor to conform to new processRow api
- getRegionsInfo: use new processRow api, and get the server address out of the RowResult, rather than relying on the one that is cached in HConnectionManager
ScannerCallable:
- make constructor public, add javadoc
HMaster
- change createTable to get the HRegionInfo out of the RowResult and compare table name from HRegionInfo with that in the HRegionInfo of the table being created, instead of parsing the table name out of the row
TestHTable
- modify test's MetaScannerVisitor to conform to the new processRow api
TestHRegionInfo
- remove testParse as parse method no longer exists
table.jsp
- change catch of IOException to catch Exception and print stack trace. At least you'll be able to see why the server crashes if it does.
git-svn-id: https://svn.apache.org/repos/asf/hadoop/hbase/trunk@678208 13f79535-47bb-0310-9956-ffa450edef68
Changed InternalScanner API:
from
{code}
boolean next(HStoreKey, SortedMap<byte[], byte>)
{code}
to
{code}
boolean next(HStoreKey, SortedMap<byte[], Cell>)
{code}
Files changed as a result of API change:
HMerge.java, ColumnValueFilter, PageRowFilter, RegExpRowFilter, RowFilterInterface, RowFilterSet, StopRowFilter, WhileMatchRowFilter, HAbstractScanner, HRegion$HScanner, HRegionServer, HStoreScanner, InternalScanner, Memcache, StoreFileScanner, MetaUtils, HBaseTestCase, TestScannerAPI, TimestampTestBase, TestRegExpRowFilter, TestRowFilterAfterWrite, TestRowFilterOnMultipleFamilies, TestRowFilterSet, TestGet2, TestHMemcache, TestHRegion, TestScanner, TestSplit
Update comments in Flusher
HRegion.internalFlushCache, HStore.flushCache, HStore.internalFlushCache now returns true only if a flush was completed and a compaction is needed.
HRegion.internalFlushCache now includes region name in DroppedSnapshotException
When creating a reader during compaction, don't bother with bloom filter since we won't use it.
StoreFileScanner locks the store for read while it is in the constructor. It also does not load the bloom filter when it opens readers on the store files.
TestScannerTimes - new regression test for HBASE-737
git-svn-id: https://svn.apache.org/repos/asf/hadoop/hbase/trunk@677113 13f79535-47bb-0310-9956-ffa450edef68
HBASE-681 NPE in Memcache
HAbstractScanner
- remove HAbstactScanner.iterator() - iterator is not a method on InternalScanner
HRegion
- make getScanner more efficient by iterating only once to find the stores we need to scan
- only pass columns relevant to a store to a HStoreScanner
- remove HScanner.iterator() - iterator is not a method on InternalScanner
Memcache, MemcacheScanner
- Fix NPE in Memcache
- never return HConstants.LATEST_TIMESTAMP as the timestamp value for a row. Instead use the largest timestamp from the cells being returned. This allows a scanner to determine a timestamp that can be used to fetch the same data again should new versions be inserted later.
StoreFileScanner
- getNextViableRow would find a row that matched the row key, but did not consider the requested timestamp. Now if the row it finds has a timestamp greater than the one desired it advances to determine if a row with a timestamp less than or equal to the requested one exists since timestamps are sorted descending.
- removed an unnecessary else
testScanMultipleVersions
- Test program that fails on current trunk but passes when this patch is applied.
git-svn-id: https://svn.apache.org/repos/asf/hadoop/hbase/trunk@670124 13f79535-47bb-0310-9956-ffa450edef68
-Change HServerLoad's getLoad method to ignore the number of requests, thus causing RegionManager to assign based merely on number of regions per server
git-svn-id: https://svn.apache.org/repos/asf/hadoop/hbase/trunk@669533 13f79535-47bb-0310-9956-ffa450edef68
