http://git-wip-us.apache.org/repos/asf/hbase-site/blob/2777c693/devapidocs/src-html/org/apache/hadoop/hbase/regionserver/StoreScanner.StoreScannerCompactionRace.html ---------------------------------------------------------------------- diff --git a/devapidocs/src-html/org/apache/hadoop/hbase/regionserver/StoreScanner.StoreScannerCompactionRace.html b/devapidocs/src-html/org/apache/hadoop/hbase/regionserver/StoreScanner.StoreScannerCompactionRace.html index b32645e..7968115 100644 --- a/devapidocs/src-html/org/apache/hadoop/hbase/regionserver/StoreScanner.StoreScannerCompactionRace.html +++ b/devapidocs/src-html/org/apache/hadoop/hbase/regionserver/StoreScanner.StoreScannerCompactionRace.html @@ -30,590 +30,590 @@ <span class="sourceLineNo">022</span>import java.io.IOException;<a name="line.22"></a> <span class="sourceLineNo">023</span>import java.io.InterruptedIOException;<a name="line.23"></a> <span class="sourceLineNo">024</span>import java.util.ArrayList;<a name="line.24"></a> -<span class="sourceLineNo">025</span>import java.util.HashMap;<a name="line.25"></a> -<span class="sourceLineNo">026</span>import java.util.List;<a name="line.26"></a> -<span class="sourceLineNo">027</span>import java.util.Map;<a name="line.27"></a> -<span class="sourceLineNo">028</span>import java.util.NavigableSet;<a name="line.28"></a> -<span class="sourceLineNo">029</span>import java.util.concurrent.CountDownLatch;<a name="line.29"></a> -<span class="sourceLineNo">030</span>import java.util.concurrent.locks.ReentrantLock;<a name="line.30"></a> -<span class="sourceLineNo">031</span><a name="line.31"></a> -<span class="sourceLineNo">032</span>import org.apache.commons.logging.Log;<a name="line.32"></a> -<span class="sourceLineNo">033</span>import org.apache.commons.logging.LogFactory;<a name="line.33"></a> -<span class="sourceLineNo">034</span>import org.apache.hadoop.hbase.Cell;<a name="line.34"></a> -<span class="sourceLineNo">035</span>import org.apache.hadoop.hbase.CellComparator;<a name="line.35"></a> -<span class="sourceLineNo">036</span>import org.apache.hadoop.hbase.CellUtil;<a name="line.36"></a> -<span class="sourceLineNo">037</span>import org.apache.hadoop.hbase.DoNotRetryIOException;<a name="line.37"></a> -<span class="sourceLineNo">038</span>import org.apache.hadoop.hbase.HConstants;<a name="line.38"></a> -<span class="sourceLineNo">039</span>import org.apache.hadoop.hbase.KeyValue;<a name="line.39"></a> -<span class="sourceLineNo">040</span>import org.apache.hadoop.hbase.KeyValueUtil;<a name="line.40"></a> -<span class="sourceLineNo">041</span>import org.apache.hadoop.hbase.classification.InterfaceAudience;<a name="line.41"></a> -<span class="sourceLineNo">042</span>import org.apache.hadoop.hbase.client.IsolationLevel;<a name="line.42"></a> -<span class="sourceLineNo">043</span>import org.apache.hadoop.hbase.client.Scan;<a name="line.43"></a> -<span class="sourceLineNo">044</span>import org.apache.hadoop.hbase.executor.ExecutorService;<a name="line.44"></a> -<span class="sourceLineNo">045</span>import org.apache.hadoop.hbase.filter.Filter;<a name="line.45"></a> -<span class="sourceLineNo">046</span>import org.apache.hadoop.hbase.regionserver.ScannerContext.LimitScope;<a name="line.46"></a> -<span class="sourceLineNo">047</span>import org.apache.hadoop.hbase.regionserver.ScannerContext.NextState;<a name="line.47"></a> -<span class="sourceLineNo">048</span>import org.apache.hadoop.hbase.regionserver.handler.ParallelSeekHandler;<a name="line.48"></a> -<span class="sourceLineNo">049</span>import org.apache.hadoop.hbase.regionserver.querymatcher.CompactionScanQueryMatcher;<a name="line.49"></a> -<span class="sourceLineNo">050</span>import org.apache.hadoop.hbase.regionserver.querymatcher.LegacyScanQueryMatcher;<a name="line.50"></a> -<span class="sourceLineNo">051</span>import org.apache.hadoop.hbase.regionserver.querymatcher.ScanQueryMatcher;<a name="line.51"></a> -<span class="sourceLineNo">052</span>import org.apache.hadoop.hbase.regionserver.querymatcher.UserScanQueryMatcher;<a name="line.52"></a> -<span class="sourceLineNo">053</span>import org.apache.hadoop.hbase.util.CollectionUtils;<a name="line.53"></a> -<span class="sourceLineNo">054</span>import org.apache.hadoop.hbase.util.EnvironmentEdgeManager;<a name="line.54"></a> +<span class="sourceLineNo">025</span>import java.util.List;<a name="line.25"></a> +<span class="sourceLineNo">026</span>import java.util.NavigableSet;<a name="line.26"></a> +<span class="sourceLineNo">027</span>import java.util.concurrent.CountDownLatch;<a name="line.27"></a> +<span class="sourceLineNo">028</span>import java.util.concurrent.locks.ReentrantLock;<a name="line.28"></a> +<span class="sourceLineNo">029</span><a name="line.29"></a> +<span class="sourceLineNo">030</span>import org.apache.commons.logging.Log;<a name="line.30"></a> +<span class="sourceLineNo">031</span>import org.apache.commons.logging.LogFactory;<a name="line.31"></a> +<span class="sourceLineNo">032</span>import org.apache.hadoop.hbase.Cell;<a name="line.32"></a> +<span class="sourceLineNo">033</span>import org.apache.hadoop.hbase.CellComparator;<a name="line.33"></a> +<span class="sourceLineNo">034</span>import org.apache.hadoop.hbase.CellUtil;<a name="line.34"></a> +<span class="sourceLineNo">035</span>import org.apache.hadoop.hbase.DoNotRetryIOException;<a name="line.35"></a> +<span class="sourceLineNo">036</span>import org.apache.hadoop.hbase.HConstants;<a name="line.36"></a> +<span class="sourceLineNo">037</span>import org.apache.hadoop.hbase.KeyValue;<a name="line.37"></a> +<span class="sourceLineNo">038</span>import org.apache.hadoop.hbase.KeyValueUtil;<a name="line.38"></a> +<span class="sourceLineNo">039</span>import org.apache.hadoop.hbase.classification.InterfaceAudience;<a name="line.39"></a> +<span class="sourceLineNo">040</span>import org.apache.hadoop.hbase.client.IsolationLevel;<a name="line.40"></a> +<span class="sourceLineNo">041</span>import org.apache.hadoop.hbase.client.Scan;<a name="line.41"></a> +<span class="sourceLineNo">042</span>import org.apache.hadoop.hbase.executor.ExecutorService;<a name="line.42"></a> +<span class="sourceLineNo">043</span>import org.apache.hadoop.hbase.filter.Filter;<a name="line.43"></a> +<span class="sourceLineNo">044</span>import org.apache.hadoop.hbase.regionserver.ScannerContext.LimitScope;<a name="line.44"></a> +<span class="sourceLineNo">045</span>import org.apache.hadoop.hbase.regionserver.ScannerContext.NextState;<a name="line.45"></a> +<span class="sourceLineNo">046</span>import org.apache.hadoop.hbase.regionserver.handler.ParallelSeekHandler;<a name="line.46"></a> +<span class="sourceLineNo">047</span>import org.apache.hadoop.hbase.regionserver.querymatcher.CompactionScanQueryMatcher;<a name="line.47"></a> +<span class="sourceLineNo">048</span>import org.apache.hadoop.hbase.regionserver.querymatcher.LegacyScanQueryMatcher;<a name="line.48"></a> +<span class="sourceLineNo">049</span>import org.apache.hadoop.hbase.regionserver.querymatcher.ScanQueryMatcher;<a name="line.49"></a> +<span class="sourceLineNo">050</span>import org.apache.hadoop.hbase.regionserver.querymatcher.UserScanQueryMatcher;<a name="line.50"></a> +<span class="sourceLineNo">051</span>import org.apache.hadoop.hbase.util.CollectionUtils;<a name="line.51"></a> +<span class="sourceLineNo">052</span>import org.apache.hadoop.hbase.util.EnvironmentEdgeManager;<a name="line.52"></a> +<span class="sourceLineNo">053</span><a name="line.53"></a> +<span class="sourceLineNo">054</span>import com.google.common.annotations.VisibleForTesting;<a name="line.54"></a> <span class="sourceLineNo">055</span><a name="line.55"></a> -<span class="sourceLineNo">056</span>import com.google.common.annotations.VisibleForTesting;<a name="line.56"></a> -<span class="sourceLineNo">057</span><a name="line.57"></a> -<span class="sourceLineNo">058</span>/**<a name="line.58"></a> -<span class="sourceLineNo">059</span> * Scanner scans both the memstore and the Store. Coalesce KeyValue stream into List&lt;KeyValue&gt;<a name="line.59"></a> -<span class="sourceLineNo">060</span> * for a single row.<a name="line.60"></a> -<span class="sourceLineNo">061</span> * <p><a name="line.61"></a> -<span class="sourceLineNo">062</span> * The implementation is not thread safe. So there will be no race between next and close. The only<a name="line.62"></a> -<span class="sourceLineNo">063</span> * exception is updateReaders, it will be called in the memstore flush thread to indicate that there<a name="line.63"></a> -<span class="sourceLineNo">064</span> * is a flush.<a name="line.64"></a> -<span class="sourceLineNo">065</span> */<a name="line.65"></a> -<span class="sourceLineNo">066</span>@InterfaceAudience.Private<a name="line.66"></a> -<span class="sourceLineNo">067</span>public class StoreScanner extends NonReversedNonLazyKeyValueScanner<a name="line.67"></a> -<span class="sourceLineNo">068</span> implements KeyValueScanner, InternalScanner, ChangedReadersObserver {<a name="line.68"></a> -<span class="sourceLineNo">069</span> private static final Log LOG = LogFactory.getLog(StoreScanner.class);<a name="line.69"></a> -<span class="sourceLineNo">070</span> // In unit tests, the store could be null<a name="line.70"></a> -<span class="sourceLineNo">071</span> protected final Store store;<a name="line.71"></a> -<span class="sourceLineNo">072</span> private ScanQueryMatcher matcher;<a name="line.72"></a> -<span class="sourceLineNo">073</span> protected KeyValueHeap heap;<a name="line.73"></a> -<span class="sourceLineNo">074</span> private boolean cacheBlocks;<a name="line.74"></a> -<span class="sourceLineNo">075</span><a name="line.75"></a> -<span class="sourceLineNo">076</span> private long countPerRow = 0;<a name="line.76"></a> -<span class="sourceLineNo">077</span> private int storeLimit = -1;<a name="line.77"></a> -<span class="sourceLineNo">078</span> private int storeOffset = 0;<a name="line.78"></a> -<span class="sourceLineNo">079</span><a name="line.79"></a> -<span class="sourceLineNo">080</span> // Used to indicate that the scanner has closed (see HBASE-1107)<a name="line.80"></a> -<span class="sourceLineNo">081</span> // Do not need to be volatile because it's always accessed via synchronized methods<a name="line.81"></a> -<span class="sourceLineNo">082</span> private boolean closing = false;<a name="line.82"></a> -<span class="sourceLineNo">083</span> private final boolean get;<a name="line.83"></a> -<span class="sourceLineNo">084</span> private final boolean explicitColumnQuery;<a name="line.84"></a> -<span class="sourceLineNo">085</span> private final boolean useRowColBloom;<a name="line.85"></a> -<span class="sourceLineNo">086</span> /**<a name="line.86"></a> -<span class="sourceLineNo">087</span> * A flag that enables StoreFileScanner parallel-seeking<a name="line.87"></a> -<span class="sourceLineNo">088</span> */<a name="line.88"></a> -<span class="sourceLineNo">089</span> private boolean parallelSeekEnabled = false;<a name="line.89"></a> -<span class="sourceLineNo">090</span> private ExecutorService executor;<a name="line.90"></a> -<span class="sourceLineNo">091</span> private final Scan scan;<a name="line.91"></a> -<span class="sourceLineNo">092</span> private final long oldestUnexpiredTS;<a name="line.92"></a> -<span class="sourceLineNo">093</span> private final long now;<a name="line.93"></a> -<span class="sourceLineNo">094</span> private final int minVersions;<a name="line.94"></a> -<span class="sourceLineNo">095</span> private final long maxRowSize;<a name="line.95"></a> -<span class="sourceLineNo">096</span> private final long cellsPerHeartbeatCheck;<a name="line.96"></a> -<span class="sourceLineNo">097</span><a name="line.97"></a> -<span class="sourceLineNo">098</span> // 1) Collects all the KVHeap that are eagerly getting closed during the<a name="line.98"></a> -<span class="sourceLineNo">099</span> // course of a scan<a name="line.99"></a> -<span class="sourceLineNo">100</span> // 2) Collects the unused memstore scanners. If we close the memstore scanners<a name="line.100"></a> -<span class="sourceLineNo">101</span> // before sending data to client, the chunk may be reclaimed by other<a name="line.101"></a> -<span class="sourceLineNo">102</span> // updates and the data will be corrupt.<a name="line.102"></a> -<span class="sourceLineNo">103</span> private final List<KeyValueScanner> scannersForDelayedClose = new ArrayList<>();<a name="line.103"></a> -<span class="sourceLineNo">104</span><a name="line.104"></a> -<span class="sourceLineNo">105</span> /**<a name="line.105"></a> -<span class="sourceLineNo">106</span> * The number of KVs seen by the scanner. Includes explicitly skipped KVs, but not<a name="line.106"></a> -<span class="sourceLineNo">107</span> * KVs skipped via seeking to next row/column. TODO: estimate them?<a name="line.107"></a> -<span class="sourceLineNo">108</span> */<a name="line.108"></a> -<span class="sourceLineNo">109</span> private long kvsScanned = 0;<a name="line.109"></a> -<span class="sourceLineNo">110</span> private Cell prevCell = null;<a name="line.110"></a> -<span class="sourceLineNo">111</span><a name="line.111"></a> -<span class="sourceLineNo">112</span> private final long preadMaxBytes;<a name="line.112"></a> -<span class="sourceLineNo">113</span> private long bytesRead;<a name="line.113"></a> -<span class="sourceLineNo">114</span><a name="line.114"></a> -<span class="sourceLineNo">115</span> /** We don't ever expect to change this, the constant is just for clarity. */<a name="line.115"></a> -<span class="sourceLineNo">116</span> static final boolean LAZY_SEEK_ENABLED_BY_DEFAULT = true;<a name="line.116"></a> -<span class="sourceLineNo">117</span> public static final String STORESCANNER_PARALLEL_SEEK_ENABLE =<a name="line.117"></a> -<span class="sourceLineNo">118</span> "hbase.storescanner.parallel.seek.enable";<a name="line.118"></a> -<span class="sourceLineNo">119</span><a name="line.119"></a> -<span class="sourceLineNo">120</span> /** Used during unit testing to ensure that lazy seek does save seek ops */<a name="line.120"></a> -<span class="sourceLineNo">121</span> private static boolean lazySeekEnabledGlobally = LAZY_SEEK_ENABLED_BY_DEFAULT;<a name="line.121"></a> -<span class="sourceLineNo">122</span><a name="line.122"></a> -<span class="sourceLineNo">123</span> /**<a name="line.123"></a> -<span class="sourceLineNo">124</span> * The number of cells scanned in between timeout checks. Specifying a larger value means that<a name="line.124"></a> -<span class="sourceLineNo">125</span> * timeout checks will occur less frequently. Specifying a small value will lead to more frequent<a name="line.125"></a> -<span class="sourceLineNo">126</span> * timeout checks.<a name="line.126"></a> -<span class="sourceLineNo">127</span> */<a name="line.127"></a> -<span class="sourceLineNo">128</span> public static final String HBASE_CELLS_SCANNED_PER_HEARTBEAT_CHECK =<a name="line.128"></a> -<span class="sourceLineNo">129</span> "hbase.cells.scanned.per.heartbeat.check";<a name="line.129"></a> -<span class="sourceLineNo">130</span><a name="line.130"></a> -<span class="sourceLineNo">131</span> /**<a name="line.131"></a> -<span class="sourceLineNo">132</span> * Default value of {@link #HBASE_CELLS_SCANNED_PER_HEARTBEAT_CHECK}.<a name="line.132"></a> -<span class="sourceLineNo">133</span> */<a name="line.133"></a> -<span class="sourceLineNo">134</span> public static final long DEFAULT_HBASE_CELLS_SCANNED_PER_HEARTBEAT_CHECK = 10000;<a name="line.134"></a> -<span class="sourceLineNo">135</span><a name="line.135"></a> -<span class="sourceLineNo">136</span> /**<a name="line.136"></a> -<span class="sourceLineNo">137</span> * If the read type if Scan.ReadType.DEFAULT, we will start with pread, and if the kvs we scanned<a name="line.137"></a> -<span class="sourceLineNo">138</span> * reaches this limit, we will reopen the scanner with stream. The default value is 4 times of<a name="line.138"></a> -<span class="sourceLineNo">139</span> * block size for this store.<a name="line.139"></a> -<span class="sourceLineNo">140</span> */<a name="line.140"></a> -<span class="sourceLineNo">141</span> public static final String STORESCANNER_PREAD_MAX_BYTES = "hbase.storescanner.pread.max.bytes";<a name="line.141"></a> +<span class="sourceLineNo">056</span>/**<a name="line.56"></a> +<span class="sourceLineNo">057</span> * Scanner scans both the memstore and the Store. Coalesce KeyValue stream into List&lt;KeyValue&gt;<a name="line.57"></a> +<span class="sourceLineNo">058</span> * for a single row.<a name="line.58"></a> +<span class="sourceLineNo">059</span> * <p><a name="line.59"></a> +<span class="sourceLineNo">060</span> * The implementation is not thread safe. So there will be no race between next and close. The only<a name="line.60"></a> +<span class="sourceLineNo">061</span> * exception is updateReaders, it will be called in the memstore flush thread to indicate that there<a name="line.61"></a> +<span class="sourceLineNo">062</span> * is a flush.<a name="line.62"></a> +<span class="sourceLineNo">063</span> */<a name="line.63"></a> +<span class="sourceLineNo">064</span>@InterfaceAudience.Private<a name="line.64"></a> +<span class="sourceLineNo">065</span>public class StoreScanner extends NonReversedNonLazyKeyValueScanner<a name="line.65"></a> +<span class="sourceLineNo">066</span> implements KeyValueScanner, InternalScanner, ChangedReadersObserver {<a name="line.66"></a> +<span class="sourceLineNo">067</span> private static final Log LOG = LogFactory.getLog(StoreScanner.class);<a name="line.67"></a> +<span class="sourceLineNo">068</span> // In unit tests, the store could be null<a name="line.68"></a> +<span class="sourceLineNo">069</span> protected final Store store;<a name="line.69"></a> +<span class="sourceLineNo">070</span> private ScanQueryMatcher matcher;<a name="line.70"></a> +<span class="sourceLineNo">071</span> protected KeyValueHeap heap;<a name="line.71"></a> +<span class="sourceLineNo">072</span> private boolean cacheBlocks;<a name="line.72"></a> +<span class="sourceLineNo">073</span><a name="line.73"></a> +<span class="sourceLineNo">074</span> private long countPerRow = 0;<a name="line.74"></a> +<span class="sourceLineNo">075</span> private int storeLimit = -1;<a name="line.75"></a> +<span class="sourceLineNo">076</span> private int storeOffset = 0;<a name="line.76"></a> +<span class="sourceLineNo">077</span><a name="line.77"></a> +<span class="sourceLineNo">078</span> // Used to indicate that the scanner has closed (see HBASE-1107)<a name="line.78"></a> +<span class="sourceLineNo">079</span> // Do not need to be volatile because it's always accessed via synchronized methods<a name="line.79"></a> +<span class="sourceLineNo">080</span> private boolean closing = false;<a name="line.80"></a> +<span class="sourceLineNo">081</span> private final boolean get;<a name="line.81"></a> +<span class="sourceLineNo">082</span> private final boolean explicitColumnQuery;<a name="line.82"></a> +<span class="sourceLineNo">083</span> private final boolean useRowColBloom;<a name="line.83"></a> +<span class="sourceLineNo">084</span> /**<a name="line.84"></a> +<span class="sourceLineNo">085</span> * A flag that enables StoreFileScanner parallel-seeking<a name="line.85"></a> +<span class="sourceLineNo">086</span> */<a name="line.86"></a> +<span class="sourceLineNo">087</span> private boolean parallelSeekEnabled = false;<a name="line.87"></a> +<span class="sourceLineNo">088</span> private ExecutorService executor;<a name="line.88"></a> +<span class="sourceLineNo">089</span> private final Scan scan;<a name="line.89"></a> +<span class="sourceLineNo">090</span> private final long oldestUnexpiredTS;<a name="line.90"></a> +<span class="sourceLineNo">091</span> private final long now;<a name="line.91"></a> +<span class="sourceLineNo">092</span> private final int minVersions;<a name="line.92"></a> +<span class="sourceLineNo">093</span> private final long maxRowSize;<a name="line.93"></a> +<span class="sourceLineNo">094</span> private final long cellsPerHeartbeatCheck;<a name="line.94"></a> +<span class="sourceLineNo">095</span><a name="line.95"></a> +<span class="sourceLineNo">096</span> // 1) Collects all the KVHeap that are eagerly getting closed during the<a name="line.96"></a> +<span class="sourceLineNo">097</span> // course of a scan<a name="line.97"></a> +<span class="sourceLineNo">098</span> // 2) Collects the unused memstore scanners. If we close the memstore scanners<a name="line.98"></a> +<span class="sourceLineNo">099</span> // before sending data to client, the chunk may be reclaimed by other<a name="line.99"></a> +<span class="sourceLineNo">100</span> // updates and the data will be corrupt.<a name="line.100"></a> +<span class="sourceLineNo">101</span> private final List<KeyValueScanner> scannersForDelayedClose = new ArrayList<>();<a name="line.101"></a> +<span class="sourceLineNo">102</span><a name="line.102"></a> +<span class="sourceLineNo">103</span> /**<a name="line.103"></a> +<span class="sourceLineNo">104</span> * The number of KVs seen by the scanner. Includes explicitly skipped KVs, but not<a name="line.104"></a> +<span class="sourceLineNo">105</span> * KVs skipped via seeking to next row/column. TODO: estimate them?<a name="line.105"></a> +<span class="sourceLineNo">106</span> */<a name="line.106"></a> +<span class="sourceLineNo">107</span> private long kvsScanned = 0;<a name="line.107"></a> +<span class="sourceLineNo">108</span> private Cell prevCell = null;<a name="line.108"></a> +<span class="sourceLineNo">109</span><a name="line.109"></a> +<span class="sourceLineNo">110</span> private final long preadMaxBytes;<a name="line.110"></a> +<span class="sourceLineNo">111</span> private long bytesRead;<a name="line.111"></a> +<span class="sourceLineNo">112</span><a name="line.112"></a> +<span class="sourceLineNo">113</span> /** We don't ever expect to change this, the constant is just for clarity. */<a name="line.113"></a> +<span class="sourceLineNo">114</span> static final boolean LAZY_SEEK_ENABLED_BY_DEFAULT = true;<a name="line.114"></a> +<span class="sourceLineNo">115</span> public static final String STORESCANNER_PARALLEL_SEEK_ENABLE =<a name="line.115"></a> +<span class="sourceLineNo">116</span> "hbase.storescanner.parallel.seek.enable";<a name="line.116"></a> +<span class="sourceLineNo">117</span><a name="line.117"></a> +<span class="sourceLineNo">118</span> /** Used during unit testing to ensure that lazy seek does save seek ops */<a name="line.118"></a> +<span class="sourceLineNo">119</span> private static boolean lazySeekEnabledGlobally = LAZY_SEEK_ENABLED_BY_DEFAULT;<a name="line.119"></a> +<span class="sourceLineNo">120</span><a name="line.120"></a> +<span class="sourceLineNo">121</span> /**<a name="line.121"></a> +<span class="sourceLineNo">122</span> * The number of cells scanned in between timeout checks. Specifying a larger value means that<a name="line.122"></a> +<span class="sourceLineNo">123</span> * timeout checks will occur less frequently. Specifying a small value will lead to more frequent<a name="line.123"></a> +<span class="sourceLineNo">124</span> * timeout checks.<a name="line.124"></a> +<span class="sourceLineNo">125</span> */<a name="line.125"></a> +<span class="sourceLineNo">126</span> public static final String HBASE_CELLS_SCANNED_PER_HEARTBEAT_CHECK =<a name="line.126"></a> +<span class="sourceLineNo">127</span> "hbase.cells.scanned.per.heartbeat.check";<a name="line.127"></a> +<span class="sourceLineNo">128</span><a name="line.128"></a> +<span class="sourceLineNo">129</span> /**<a name="line.129"></a> +<span class="sourceLineNo">130</span> * Default value of {@link #HBASE_CELLS_SCANNED_PER_HEARTBEAT_CHECK}.<a name="line.130"></a> +<span class="sourceLineNo">131</span> */<a name="line.131"></a> +<span class="sourceLineNo">132</span> public static final long DEFAULT_HBASE_CELLS_SCANNED_PER_HEARTBEAT_CHECK = 10000;<a name="line.132"></a> +<span class="sourceLineNo">133</span><a name="line.133"></a> +<span class="sourceLineNo">134</span> /**<a name="line.134"></a> +<span class="sourceLineNo">135</span> * If the read type if Scan.ReadType.DEFAULT, we will start with pread, and if the kvs we scanned<a name="line.135"></a> +<span class="sourceLineNo">136</span> * reaches this limit, we will reopen the scanner with stream. The default value is 4 times of<a name="line.136"></a> +<span class="sourceLineNo">137</span> * block size for this store.<a name="line.137"></a> +<span class="sourceLineNo">138</span> */<a name="line.138"></a> +<span class="sourceLineNo">139</span> public static final String STORESCANNER_PREAD_MAX_BYTES = "hbase.storescanner.pread.max.bytes";<a name="line.139"></a> +<span class="sourceLineNo">140</span><a name="line.140"></a> +<span class="sourceLineNo">141</span> private final Scan.ReadType readType;<a name="line.141"></a> <span class="sourceLineNo">142</span><a name="line.142"></a> -<span class="sourceLineNo">143</span> private final Scan.ReadType readType;<a name="line.143"></a> -<span class="sourceLineNo">144</span><a name="line.144"></a> -<span class="sourceLineNo">145</span> // A flag whether use pread for scan<a name="line.145"></a> -<span class="sourceLineNo">146</span> // it maybe changed if we use Scan.ReadType.DEFAULT and we have read lots of data.<a name="line.146"></a> -<span class="sourceLineNo">147</span> private boolean scanUsePread;<a name="line.147"></a> -<span class="sourceLineNo">148</span> // Indicates whether there was flush during the course of the scan<a name="line.148"></a> -<span class="sourceLineNo">149</span> private volatile boolean flushed = false;<a name="line.149"></a> -<span class="sourceLineNo">150</span> // generally we get one file from a flush<a name="line.150"></a> -<span class="sourceLineNo">151</span> private final List<StoreFile> flushedStoreFiles = new ArrayList<>(1);<a name="line.151"></a> -<span class="sourceLineNo">152</span> // Since CompactingMemstore is now default, we get three memstore scanners from a flush<a name="line.152"></a> -<span class="sourceLineNo">153</span> private final List<KeyValueScanner> memStoreScannersAfterFlush = new ArrayList<>(3);<a name="line.153"></a> -<span class="sourceLineNo">154</span> // The current list of scanners<a name="line.154"></a> -<span class="sourceLineNo">155</span> @VisibleForTesting<a name="line.155"></a> -<span class="sourceLineNo">156</span> final List<KeyValueScanner> currentScanners = new ArrayList<>();<a name="line.156"></a> -<span class="sourceLineNo">157</span> // flush update lock<a name="line.157"></a> -<span class="sourceLineNo">158</span> private final ReentrantLock flushLock = new ReentrantLock();<a name="line.158"></a> -<span class="sourceLineNo">159</span><a name="line.159"></a> -<span class="sourceLineNo">160</span> protected final long readPt;<a name="line.160"></a> -<span class="sourceLineNo">161</span><a name="line.161"></a> -<span class="sourceLineNo">162</span> // used by the injection framework to test race between StoreScanner construction and compaction<a name="line.162"></a> -<span class="sourceLineNo">163</span> enum StoreScannerCompactionRace {<a name="line.163"></a> -<span class="sourceLineNo">164</span> BEFORE_SEEK,<a name="line.164"></a> -<span class="sourceLineNo">165</span> AFTER_SEEK,<a name="line.165"></a> -<span class="sourceLineNo">166</span> COMPACT_COMPLETE<a name="line.166"></a> -<span class="sourceLineNo">167</span> }<a name="line.167"></a> -<span class="sourceLineNo">168</span><a name="line.168"></a> -<span class="sourceLineNo">169</span> /** An internal constructor. */<a name="line.169"></a> -<span class="sourceLineNo">170</span> protected StoreScanner(Store store, Scan scan, final ScanInfo scanInfo,<a name="line.170"></a> -<span class="sourceLineNo">171</span> final NavigableSet<byte[]> columns, long readPt, boolean cacheBlocks, ScanType scanType) {<a name="line.171"></a> -<span class="sourceLineNo">172</span> this.readPt = readPt;<a name="line.172"></a> -<span class="sourceLineNo">173</span> this.store = store;<a name="line.173"></a> -<span class="sourceLineNo">174</span> this.cacheBlocks = cacheBlocks;<a name="line.174"></a> -<span class="sourceLineNo">175</span> get = scan.isGetScan();<a name="line.175"></a> -<span class="sourceLineNo">176</span> int numCol = columns == null ? 0 : columns.size();<a name="line.176"></a> -<span class="sourceLineNo">177</span> explicitColumnQuery = numCol > 0;<a name="line.177"></a> -<span class="sourceLineNo">178</span> this.scan = scan;<a name="line.178"></a> -<span class="sourceLineNo">179</span> this.now = EnvironmentEdgeManager.currentTime();<a name="line.179"></a> -<span class="sourceLineNo">180</span> this.oldestUnexpiredTS = scan.isRaw() ? 0L : now - scanInfo.getTtl();<a name="line.180"></a> -<span class="sourceLineNo">181</span> this.minVersions = scanInfo.getMinVersions();<a name="line.181"></a> -<span class="sourceLineNo">182</span><a name="line.182"></a> -<span class="sourceLineNo">183</span> // We look up row-column Bloom filters for multi-column queries as part of<a name="line.183"></a> -<span class="sourceLineNo">184</span> // the seek operation. However, we also look the row-column Bloom filter<a name="line.184"></a> -<span class="sourceLineNo">185</span> // for multi-row (non-"get") scans because this is not done in<a name="line.185"></a> -<span class="sourceLineNo">186</span> // StoreFile.passesBloomFilter(Scan, SortedSet<byte[]>).<a name="line.186"></a> -<span class="sourceLineNo">187</span> this.useRowColBloom = numCol > 1 || (!get && numCol == 1);<a name="line.187"></a> -<span class="sourceLineNo">188</span><a name="line.188"></a> -<span class="sourceLineNo">189</span> this.maxRowSize = scanInfo.getTableMaxRowSize();<a name="line.189"></a> -<span class="sourceLineNo">190</span> if (get) {<a name="line.190"></a> -<span class="sourceLineNo">191</span> this.readType = Scan.ReadType.PREAD;<a name="line.191"></a> -<span class="sourceLineNo">192</span> this.scanUsePread = true;<a name="line.192"></a> -<span class="sourceLineNo">193</span> } else if(scanType != scanType.USER_SCAN) {<a name="line.193"></a> -<span class="sourceLineNo">194</span> // For compaction scanners never use Pread as already we have stream based scanners on the<a name="line.194"></a> -<span class="sourceLineNo">195</span> // store files to be compacted<a name="line.195"></a> -<span class="sourceLineNo">196</span> this.readType = Scan.ReadType.STREAM;<a name="line.196"></a> -<span class="sourceLineNo">197</span> this.scanUsePread = false;<a name="line.197"></a> -<span class="sourceLineNo">198</span> } else {<a name="line.198"></a> -<span class="sourceLineNo">199</span> if (scan.getReadType() == Scan.ReadType.DEFAULT) {<a name="line.199"></a> -<span class="sourceLineNo">200</span> this.readType = scanInfo.isUsePread() ? Scan.ReadType.PREAD : Scan.ReadType.DEFAULT;<a name="line.200"></a> -<span class="sourceLineNo">201</span> } else {<a name="line.201"></a> -<span class="sourceLineNo">202</span> this.readType = scan.getReadType();<a name="line.202"></a> -<span class="sourceLineNo">203</span> }<a name="line.203"></a> -<span class="sourceLineNo">204</span> // Always start with pread unless user specific stream. Will change to stream later if<a name="line.204"></a> -<span class="sourceLineNo">205</span> // readType is default if the scan keeps running for a long time.<a name="line.205"></a> -<span class="sourceLineNo">206</span> this.scanUsePread = this.readType != Scan.ReadType.STREAM;<a name="line.206"></a> -<span class="sourceLineNo">207</span> }<a name="line.207"></a> -<span class="sourceLineNo">208</span> this.preadMaxBytes = scanInfo.getPreadMaxBytes();<a name="line.208"></a> -<span class="sourceLineNo">209</span> this.cellsPerHeartbeatCheck = scanInfo.getCellsPerTimeoutCheck();<a name="line.209"></a> -<span class="sourceLineNo">210</span> // Parallel seeking is on if the config allows and more there is more than one store file.<a name="line.210"></a> -<span class="sourceLineNo">211</span> if (this.store != null && this.store.getStorefilesCount() > 1) {<a name="line.211"></a> -<span class="sourceLineNo">212</span> RegionServerServices rsService = ((HStore) store).getHRegion().getRegionServerServices();<a name="line.212"></a> -<span class="sourceLineNo">213</span> if (rsService != null && scanInfo.isParallelSeekEnabled()) {<a name="line.213"></a> -<span class="sourceLineNo">214</span> this.parallelSeekEnabled = true;<a name="line.214"></a> -<span class="sourceLineNo">215</span> this.executor = rsService.getExecutorService();<a name="line.215"></a> -<span class="sourceLineNo">216</span> }<a name="line.216"></a> -<span class="sourceLineNo">217</span> }<a name="line.217"></a> -<span class="sourceLineNo">218</span> }<a name="line.218"></a> -<span class="sourceLineNo">219</span><a name="line.219"></a> -<span class="sourceLineNo">220</span> private void addCurrentScanners(List<? extends KeyValueScanner> scanners) {<a name="line.220"></a> -<span class="sourceLineNo">221</span> this.currentScanners.addAll(scanners);<a name="line.221"></a> -<span class="sourceLineNo">222</span> }<a name="line.222"></a> -<span class="sourceLineNo">223</span><a name="line.223"></a> -<span class="sourceLineNo">224</span> /**<a name="line.224"></a> -<span class="sourceLineNo">225</span> * Opens a scanner across memstore, snapshot, and all StoreFiles. Assumes we<a name="line.225"></a> -<span class="sourceLineNo">226</span> * are not in a compaction.<a name="line.226"></a> -<span class="sourceLineNo">227</span> *<a name="line.227"></a> -<span class="sourceLineNo">228</span> * @param store who we scan<a name="line.228"></a> -<span class="sourceLineNo">229</span> * @param scan the spec<a name="line.229"></a> -<span class="sourceLineNo">230</span> * @param columns which columns we are scanning<a name="line.230"></a> -<span class="sourceLineNo">231</span> * @throws IOException<a name="line.231"></a> -<span class="sourceLineNo">232</span> */<a name="line.232"></a> -<span class="sourceLineNo">233</span> public StoreScanner(Store store, ScanInfo scanInfo, Scan scan, final NavigableSet<byte[]> columns,<a name="line.233"></a> -<span class="sourceLineNo">234</span> long readPt)<a name="line.234"></a> -<span class="sourceLineNo">235</span> throws IOException {<a name="line.235"></a> -<span class="sourceLineNo">236</span> this(store, scan, scanInfo, columns, readPt, scan.getCacheBlocks(), ScanType.USER_SCAN);<a name="line.236"></a> -<span class="sourceLineNo">237</span> if (columns != null && scan.isRaw()) {<a name="line.237"></a> -<span class="sourceLineNo">238</span> throw new DoNotRetryIOException("Cannot specify any column for a raw scan");<a name="line.238"></a> -<span class="sourceLineNo">239</span> }<a name="line.239"></a> -<span class="sourceLineNo">240</span> matcher = UserScanQueryMatcher.create(scan, scanInfo, columns, oldestUnexpiredTS, now,<a name="line.240"></a> -<span class="sourceLineNo">241</span> store.getCoprocessorHost());<a name="line.241"></a> -<span class="sourceLineNo">242</span><a name="line.242"></a> -<span class="sourceLineNo">243</span> this.store.addChangedReaderObserver(this);<a name="line.243"></a> -<span class="sourceLineNo">244</span><a name="line.244"></a> -<span class="sourceLineNo">245</span> try {<a name="line.245"></a> -<span class="sourceLineNo">246</span> // Pass columns to try to filter out unnecessary StoreFiles.<a name="line.246"></a> -<span class="sourceLineNo">247</span> List<KeyValueScanner> scanners = getScannersNoCompaction();<a name="line.247"></a> -<span class="sourceLineNo">248</span><a name="line.248"></a> -<span class="sourceLineNo">249</span> // Seek all scanners to the start of the Row (or if the exact matching row<a name="line.249"></a> -<span class="sourceLineNo">250</span> // key does not exist, then to the start of the next matching Row).<a name="line.250"></a> -<span class="sourceLineNo">251</span> // Always check bloom filter to optimize the top row seek for delete<a name="line.251"></a> -<span class="sourceLineNo">252</span> // family marker.<a name="line.252"></a> -<span class="sourceLineNo">253</span> seekScanners(scanners, matcher.getStartKey(), explicitColumnQuery && lazySeekEnabledGlobally,<a name="line.253"></a> -<span class="sourceLineNo">254</span> parallelSeekEnabled);<a name="line.254"></a> -<span class="sourceLineNo">255</span><a name="line.255"></a> -<span class="sourceLineNo">256</span> // set storeLimit<a name="line.256"></a> -<span class="sourceLineNo">257</span> this.storeLimit = scan.getMaxResultsPerColumnFamily();<a name="line.257"></a> -<span class="sourceLineNo">258</span><a name="line.258"></a> -<span class="sourceLineNo">259</span> // set rowOffset<a name="line.259"></a> -<span class="sourceLineNo">260</span> this.storeOffset = scan.getRowOffsetPerColumnFamily();<a name="line.260"></a> -<span class="sourceLineNo">261</span> addCurrentScanners(scanners);<a name="line.261"></a> -<span class="sourceLineNo">262</span> // Combine all seeked scanners with a heap<a name="line.262"></a> -<span class="sourceLineNo">263</span> resetKVHeap(scanners, store.getComparator());<a name="line.263"></a> -<span class="sourceLineNo">264</span> } catch (IOException e) {<a name="line.264"></a> -<span class="sourceLineNo">265</span> // remove us from the HStore#changedReaderObservers here or we'll have no chance to<a name="line.265"></a> -<span class="sourceLineNo">266</span> // and might cause memory leak<a name="line.266"></a> -<span class="sourceLineNo">267</span> this.store.deleteChangedReaderObserver(this);<a name="line.267"></a> -<span class="sourceLineNo">268</span> throw e;<a name="line.268"></a> -<span class="sourceLineNo">269</span> }<a name="line.269"></a> -<span class="sourceLineNo">270</span> }<a name="line.270"></a> -<span class="sourceLineNo">271</span><a name="line.271"></a> -<span class="sourceLineNo">272</span> /**<a name="line.272"></a> -<span class="sourceLineNo">273</span> * Used for compactions.<p><a name="line.273"></a> -<span class="sourceLineNo">274</span> *<a name="line.274"></a> -<span class="sourceLineNo">275</span> * Opens a scanner across specified StoreFiles.<a name="line.275"></a> -<span class="sourceLineNo">276</span> * @param store who we scan<a name="line.276"></a> -<span class="sourceLineNo">277</span> * @param scan the spec<a name="line.277"></a> -<span class="sourceLineNo">278</span> * @param scanners ancillary scanners<a name="line.278"></a> -<span class="sourceLineNo">279</span> * @param smallestReadPoint the readPoint that we should use for tracking<a name="line.279"></a> -<span class="sourceLineNo">280</span> * versions<a name="line.280"></a> -<span class="sourceLineNo">281</span> */<a name="line.281"></a> -<span class="sourceLineNo">282</span> public StoreScanner(Store store, ScanInfo scanInfo, Scan scan,<a name="line.282"></a> -<span class="sourceLineNo">283</span> List<? extends KeyValueScanner> scanners, ScanType scanType,<a name="line.283"></a> -<span class="sourceLineNo">284</span> long smallestReadPoint, long earliestPutTs) throws IOException {<a name="line.284"></a> -<span class="sourceLineNo">285</span> this(store, scanInfo, scan, scanners, scanType, smallestReadPoint, earliestPutTs, null, null);<a name="line.285"></a> -<span class="sourceLineNo">286</span> }<a name="line.286"></a> -<span class="sourceLineNo">287</span><a name="line.287"></a> -<span class="sourceLineNo">288</span> /**<a name="line.288"></a> -<span class="sourceLineNo">289</span> * Used for compactions that drop deletes from a limited range of rows.<p><a name="line.289"></a> -<span class="sourceLineNo">290</span> *<a name="line.290"></a> -<span class="sourceLineNo">291</span> * Opens a scanner across specified StoreFiles.<a name="line.291"></a> -<span class="sourceLineNo">292</span> * @param store who we scan<a name="line.292"></a> -<span class="sourceLineNo">293</span> * @param scan the spec<a name="line.293"></a> -<span class="sourceLineNo">294</span> * @param scanners ancillary scanners<a name="line.294"></a> -<span class="sourceLineNo">295</span> * @param smallestReadPoint the readPoint that we should use for tracking versions<a name="line.295"></a> -<span class="sourceLineNo">296</span> * @param dropDeletesFromRow The inclusive left bound of the range; can be EMPTY_START_ROW.<a name="line.296"></a> -<span class="sourceLineNo">297</span> * @param dropDeletesToRow The exclusive right bound of the range; can be EMPTY_END_ROW.<a name="line.297"></a> -<span class="sourceLineNo">298</span> */<a name="line.298"></a> -<span class="sourceLineNo">299</span> public StoreScanner(Store store, ScanInfo scanInfo, Scan scan,<a name="line.299"></a> -<span class="sourceLineNo">300</span> List<? extends KeyValueScanner> scanners, long smallestReadPoint, long earliestPutTs,<a name="line.300"></a> -<span class="sourceLineNo">301</span> byte[] dropDeletesFromRow, byte[] dropDeletesToRow) throws IOException {<a name="line.301"></a> -<span class="sourceLineNo">302</span> this(store, scanInfo, scan, scanners, ScanType.COMPACT_RETAIN_DELETES, smallestReadPoint,<a name="line.302"></a> -<span class="sourceLineNo">303</span> earliestPutTs, dropDeletesFromRow, dropDeletesToRow);<a name="line.303"></a> -<span class="sourceLineNo">304</span> }<a name="line.304"></a> -<span class="sourceLineNo">305</span><a name="line.305"></a> -<span class="sourceLineNo">306</span> private StoreScanner(Store store, ScanInfo scanInfo, Scan scan,<a name="line.306"></a> -<span class="sourceLineNo">307</span> List<? extends KeyValueScanner> scanners, ScanType scanType, long smallestReadPoint,<a name="line.307"></a> -<span class="sourceLineNo">308</span> long earliestPutTs, byte[] dropDeletesFromRow, byte[] dropDeletesToRow) throws IOException {<a name="line.308"></a> -<span class="sourceLineNo">309</span> this(store, scan, scanInfo, null,<a name="line.309"></a> -<span class="sourceLineNo">310</span> ((HStore) store).getHRegion().getReadPoint(IsolationLevel.READ_COMMITTED), false, scanType);<a name="line.310"></a> -<span class="sourceLineNo">311</span> if (scan.hasFilter() || (scan.getStartRow() != null && scan.getStartRow().length > 0)<a name="line.311"></a> -<span class="sourceLineNo">312</span> || (scan.getStopRow() != null && scan.getStopRow().length > 0)<a name="line.312"></a> -<span class="sourceLineNo">313</span> || !scan.getTimeRange().isAllTime()) {<a name="line.313"></a> -<span class="sourceLineNo">314</span> // use legacy query matcher since we do not consider the scan object in our code. Only used to<a name="line.314"></a> -<span class="sourceLineNo">315</span> // keep compatibility for coprocessor.<a name="line.315"></a> -<span class="sourceLineNo">316</span> matcher = LegacyScanQueryMatcher.create(scan, scanInfo, null, scanType, smallestReadPoint,<a name="line.316"></a> -<span class="sourceLineNo">317</span> earliestPutTs, oldestUnexpiredTS, now, dropDeletesFromRow, dropDeletesToRow,<a name="line.317"></a> -<span class="sourceLineNo">318</span> store.getCoprocessorHost());<a name="line.318"></a> -<span class="sourceLineNo">319</span> } else {<a name="line.319"></a> -<span class="sourceLineNo">320</span> matcher = CompactionScanQueryMatcher.create(scanInfo, scanType, smallestReadPoint,<a name="line.320"></a> -<span class="sourceLineNo">321</span> earliestPutTs, oldestUnexpiredTS, now, dropDeletesFromRow, dropDeletesToRow,<a name="line.321"></a> -<span class="sourceLineNo">322</span> store.getCoprocessorHost());<a name="line.322"></a> -<span class="sourceLineNo">323</span> }<a name="line.323"></a> -<span class="sourceLineNo">324</span><a name="line.324"></a> -<span class="sourceLineNo">325</span> // Filter the list of scanners using Bloom filters, time range, TTL, etc.<a name="line.325"></a> -<span class="sourceLineNo">326</span> scanners = selectScannersFrom(scanners);<a name="line.326"></a> -<span class="sourceLineNo">327</span><a name="line.327"></a> -<span class="sourceLineNo">328</span> // Seek all scanners to the initial key<a name="line.328"></a> -<span class="sourceLineNo">329</span> seekScanners(scanners, matcher.getStartKey(), false, parallelSeekEnabled);<a name="line.329"></a> -<span class="sourceLineNo">330</span> addCurrentScanners(scanners);<a name="line.330"></a> -<span class="sourceLineNo">331</span> // Combine all seeked scanners with a heap<a name="line.331"></a> -<span class="sourceLineNo">332</span> resetKVHeap(scanners, store.getComparator());<a name="line.332"></a> -<span class="sourceLineNo">333</span> }<a name="line.333"></a> -<span class="sourceLineNo">334</span><a name="line.334"></a> -<span class="sourceLineNo">335</span> @VisibleForTesting<a name="line.335"></a> -<span class="sourceLineNo">336</span> StoreScanner(final Scan scan, ScanInfo scanInfo,<a name="line.336"></a> -<span class="sourceLineNo">337</span> ScanType scanType, final NavigableSet<byte[]> columns,<a name="line.337"></a> -<span class="sourceLineNo">338</span> final List<? extends KeyValueScanner> scanners) throws IOException {<a name="line.338"></a> -<span class="sourceLineNo">339</span> this(scan, scanInfo, scanType, columns, scanners,<a name="line.339"></a> -<span class="sourceLineNo">340</span> HConstants.LATEST_TIMESTAMP,<a name="line.340"></a> -<span class="sourceLineNo">341</span> // 0 is passed as readpoint because the test bypasses Store<a name="line.341"></a> -<span class="sourceLineNo">342</span> 0);<a name="line.342"></a> -<span class="sourceLineNo">343</span> }<a name="line.343"></a> -<span class="sourceLineNo">344</span><a name="line.344"></a> -<span class="sourceLineNo">345</span> @VisibleForTesting<a name="line.345"></a> -<span class="sourceLineNo">346</span> StoreScanner(final Scan scan, ScanInfo scanInfo,<a name="line.346"></a> -<span class="sourceLineNo">347</span> ScanType scanType, final NavigableSet<byte[]> columns,<a name="line.347"></a> -<span class="sourceLineNo">348</span> final List<? extends KeyValueScanner> scanners, long earliestPutTs)<a name="line.348"></a> -<span class="sourceLineNo">349</span> throws IOException {<a name="line.349"></a> -<span class="sourceLineNo">350</span> this(scan, scanInfo, scanType, columns, scanners, earliestPutTs,<a name="line.350"></a> -<span class="sourceLineNo">351</span> // 0 is passed as readpoint because the test bypasses Store<a name="line.351"></a> -<span class="sourceLineNo">352</span> 0);<a name="line.352"></a> -<span class="sourceLineNo">353</span> }<a name="line.353"></a> -<span class="sourceLineNo">354</span><a name="line.354"></a> -<span class="sourceLineNo">355</span> public StoreScanner(final Scan scan, ScanInfo scanInfo, ScanType scanType,<a name="line.355"></a> -<span class="sourceLineNo">356</span> final NavigableSet<byte[]> columns, final List<? extends KeyValueScanner> scanners, long earliestPutTs,<a name="line.356"></a> -<span class="sourceLineNo">357</span> long readPt) throws IOException {<a name="line.357"></a> -<span class="sourceLineNo">358</span> this(null, scan, scanInfo, columns, readPt,<a name="line.358"></a> -<span class="sourceLineNo">359</span> scanType == ScanType.USER_SCAN ? scan.getCacheBlocks() : false, scanType);<a name="line.359"></a> -<span class="sourceLineNo">360</span> if (scanType == ScanType.USER_SCAN) {<a name="line.360"></a> -<span class="sourceLineNo">361</span> this.matcher = UserScanQueryMatcher.create(scan, scanInfo, columns, oldestUnexpiredTS, now,<a name="line.361"></a> -<span class="sourceLineNo">362</span> null);<a name="line.362"></a> -<span class="sourceLineNo">363</span> } else {<a name="line.363"></a> -<span class="sourceLineNo">364</span> if (scan.hasFilter() || (scan.getStartRow() != null && scan.getStartRow().length > 0)<a name="line.364"></a> -<span class="sourceLineNo">365</span> || (scan.getStopRow() != null && scan.getStopRow().length > 0)<a name="line.365"></a> -<span class="sourceLineNo">366</span> || !scan.getTimeRange().isAllTime() || columns != null) {<a name="line.366"></a> -<span class="sourceLineNo">367</span> // use legacy query matcher since we do not consider the scan object in our code. Only used<a name="line.367"></a> -<span class="sourceLineNo">368</span> // to keep compatibility for coprocessor.<a name="line.368"></a> -<span class="sourceLineNo">369</span> matcher = LegacyScanQueryMatcher.create(scan, scanInfo, columns, scanType, Long.MAX_VALUE,<a name="line.369"></a> -<span class="sourceLineNo">370</span> earliestPutTs, oldestUnexpiredTS, now, null, null, store.getCoprocessorHost());<a name="line.370"></a> -<span class="sourceLineNo">371</span> } else {<a name="line.371"></a> -<span class="sourceLineNo">372</span> this.matcher = CompactionScanQueryMatcher.create(scanInfo, scanType, Long.MAX_VALUE,<a name="line.372"></a> -<span class="sourceLineNo">373</span> earliestPutTs, oldestUnexpiredTS, now, null, null, null);<a name="line.373"></a> -<span class="sourceLineNo">374</span> }<a name="line.374"></a> -<span class="sourceLineNo">375</span> }<a name="line.375"></a> -<span class="sourceLineNo">376</span><a name="line.376"></a> -<span class="sourceLineNo">377</span> // Seek all scanners to the initial key<a name="line.377"></a> -<span class="sourceLineNo">378</span> seekScanners(scanners, matcher.getStartKey(), false, parallelSeekEnabled);<a name="line.378"></a> -<span class="sourceLineNo">379</span> addCurrentScanners(scanners);<a name="line.379"></a> -<span class="sourceLineNo">380</span> resetKVHeap(scanners, scanInfo.getComparator());<a name="line.380"></a> -<span class="sourceLineNo">381</span> }<a name="line.381"></a> -<span class="sourceLineNo">382</span><a name="line.382"></a> -<span class="sourceLineNo">383</span> /**<a name="line.383"></a> -<span class="sourceLineNo">384</span> * Get a filtered list of scanners. Assumes we are not in a compaction.<a name="line.384"></a> -<span class="sourceLineNo">385</span> * @return list of scanners to seek<a name="line.385"></a> -<span class="sourceLineNo">386</span> */<a name="line.386"></a> -<span class="sourceLineNo">387</span> private List<KeyValueScanner> getScannersNoCompaction() throws IOException {<a name="line.387"></a> -<span class="sourceLineNo">388</span> return selectScannersFrom(<a name="line.388"></a> -<span class="sourceLineNo">389</span> store.getScanners(cacheBlocks, scanUsePread, false, matcher, scan.getStartRow(),<a name="line.389"></a> -<span class="sourceLineNo">390</span> scan.includeStartRow(), scan.getStopRow(), scan.includeStopRow(), this.readPt));<a name="line.390"></a> -<span class="sourceLineNo">391</span> }<a name="line.391"></a> -<span class="sourceLineNo">392</span><a name="line.392"></a> -<span class="sourceLineNo">393</span> @VisibleForTesting<a name="line.393"></a> -<span class="sourceLineNo">394</span> boolean isScanUsePread() {<a name="line.394"></a> -<span class="sourceLineNo">395</span> return this.scanUsePread;<a name="line.395"></a> -<span class="sourceLineNo">396</span> }<a name="line.396"></a> -<span class="sourceLineNo">397</span> /**<a name="line.397"></a> -<span class="sourceLineNo">398</span> * Seek the specified scanners with the given key<a name="line.398"></a> -<span class="sourceLineNo">399</span> * @param scanners<a name="line.399"></a> -<span class="sourceLineNo">400</span> * @param seekKey<a name="line.400"></a> -<span class="sourceLineNo">401</span> * @param isLazy true if using lazy seek<a name="line.401"></a> -<span class="sourceLineNo">402</span> * @param isParallelSeek true if using parallel seek<a name="line.402"></a> -<span class="sourceLineNo">403</span> * @throws IOException<a name="line.403"></a> -<span class="sourceLineNo">404</span> */<a name="line.404"></a> -<span class="sourceLineNo">405</span> protected void seekScanners(List<? extends KeyValueScanner> scanners,<a name="line.405"></a> -<span class="sourceLineNo">406</span> Cell seekKey, boolean isLazy, boolean isParallelSeek)<a name="line.406"></a> -<span class="sourceLineNo">407</span> throws IOException {<a name="line.407"></a> -<span class="sourceLineNo">408</span> // Seek all scanners to the start of the Row (or if the exact matching row<a name="line.408"></a> -<span class="sourceLineNo">409</span> // key does not exist, then to the start of the next matching Row).<a name="line.409"></a> -<span class="sourceLineNo">410</span> // Always check bloom filter to optimize the top row seek for delete<a name="line.410"></a> -<span class="sourceLineNo">411</span> // family marker.<a name="line.411"></a> -<span class="sourceLineNo">412</span> if (isLazy) {<a name="line.412"></a> -<span class="sourceLineNo">413</span> for (KeyValueScanner scanner : scanners) {<a name="line.413"></a> -<span class="sourceLineNo">414</span> scanner.requestSeek(seekKey, false, true);<a name="line.414"></a> -<span class="sourceLineNo">415</span> }<a name="line.415"></a> -<span class="sourceLineNo">416</span> } else {<a name="line.416"></a> -<span class="sourceLineNo">417</span> if (!isParallelSeek) {<a name="line.417"></a> -<span class="sourceLineNo">418</span> long totalScannersSoughtBytes = 0;<a name="line.418"></a> -<span class="sourceLineNo">419</span> for (KeyValueScanner scanner : scanners) {<a name="line.419"></a> -<span class="sourceLineNo">420</span> if (matcher.isUserScan() && totalScannersSoughtBytes >= maxRowSize) {<a name="line.420"></a> -<span class="sourceLineNo">421</span> throw new RowTooBigException("Max row size allowed: " + maxRowSize<a name="line.421"></a> -<span class="sourceLineNo">422</span> + ", but row is bigger than that");<a name="line.422"></a> -<span class="sourceLineNo">423</span> }<a name="line.423"></a> -<span class="sourceLineNo">424</span> scanner.seek(seekKey);<a name="line.424"></a> -<span class="sourceLineNo">425</span> Cell c = scanner.peek();<a name="line.425"></a> -<span class="sourceLineNo">426</span> if (c != null) {<a name="line.426"></a> -<span class="sourceLineNo">427</span> totalScannersSoughtBytes += CellUtil.estimatedSerializedSizeOf(c);<a name="line.427"></a> -<span class="sourceLineNo">428</span> }<a name="line.428"></a> -<span class="sourceLineNo">429</span> }<a name="line.429"></a> -<span class="sourceLineNo">430</span> } else {<a name="line.430"></a> -<span class="sourceLineNo">431</span> parallelSeek(scanners, seekKey);<a name="line.431"></a> -<span class="sourceLineNo">432</span> }<a name="line.432"></a> -<span class="sourceLineNo">433</span> }<a name="line.433"></a> -<span class="sourceLineNo">434</span> }<a name="line.434"></a> -<span class="sourceLineNo">435</span><a name="line.435"></a> -<span class="sourceLineNo">436</span> protected void resetKVHeap(List<? extends KeyValueScanner> scanners,<a name="line.436"></a> -<span class="sourceLineNo">437</span> CellComparator comparator) throws IOException {<a name="line.437"></a> -<span class="sourceLineNo">438</span> // Combine all seeked scanners with a heap<a name="line.438"></a> -<span class="sourceLineNo">439</span> heap = new KeyValueHeap(scanners, comparator);<a name="line.439"></a> -<span class="sourceLineNo">440</span> }<a name="line.440"></a> -<span class="sourceLineNo">441</span><a name="line.441"></a> -<span class="sourceLineNo">442</span> /**<a name="line.442"></a> -<span class="sourceLineNo">443</span> * Filters the given list of scanners using Bloom filter, time range, and<a name="line.443"></a> -<span class="sourceLineNo">444</span> * TTL.<a name="line.444"></a> -<span class="sourceLineNo">445</span> * <p><a name="line.445"></a> -<span class="sourceLineNo">446</span> * Will be overridden by testcase so declared as protected.<a name="line.446"></a> -<span class="sourceLineNo">447</span> */<a name="line.447"></a> -<span class="sourceLineNo">448</span> @VisibleForTesting<a name="line.448"></a> -<span class="sourceLineNo">449</span> protected List<KeyValueScanner> selectScannersFrom(<a name="line.449"></a> -<span class="sourceLineNo">450</span> final List<? extends KeyValueScanner> allScanners) {<a name="line.450"></a> -<span class="sourceLineNo">451</span> boolean memOnly;<a name="line.451"></a> -<span class="sourceLineNo">452</span> boolean filesOnly;<a name="line.452"></a> -<span class="sourceLineNo">453</span> if (scan instanceof InternalScan) {<a name="line.453"></a> -<span class="sourceLineNo">454</span> InternalScan iscan = (InternalScan)scan;<a name="line.454"></a> -<span class="sourceLineNo">455</span> memOnly = iscan.isCheckOnlyMemStore();<a name="line.455"></a> -<span class="sourceLineNo">456</span> filesOnly = iscan.isCheckOnlyStoreFiles();<a name="line.456"></a> -<span class="sourceLineNo">457</span> } else {<a name="line.457"></a> -<span class="sourceLineNo">458</span> memOnly = false;<a name="line.458"></a> -<span class="sourceLineNo">459</span> filesOnly = false;<a name="line.459"></a> -<span class="sourceLineNo">460</span> }<a name="line.460"></a> -<span class="sourceLineNo">461</span><a name="line.461"></a> -<span class="sourceLineNo">462</span> List<KeyValueScanner> scanners = new ArrayList<>(allScanners.size());<a name="line.462"></a> -<span class="sourceLineNo">463</span><a name="line.463"></a> -<span class="sourceLineNo">464</span> // We can only exclude store files based on TTL if minVersions is set to 0.<a name="line.464"></a> -<span class="sourceLineNo">465</span> // Otherwise, we might have to return KVs that have technically expired.<a name="line.465"></a> -<span class="sourceLineNo">466</span> long expiredTimestampCutoff = minVersions == 0 ? oldestUnexpiredTS: Long.MIN_VALUE;<a name="line.466"></a> -<span class="sourceLineNo">467</span><a name="line.467"></a> -<span class="sourceLineNo">468</span> // include only those scan files which pass all filters<a name="line.468"></a> -<span class="sourceLineNo">469</span> for (KeyValueScanner kvs : allScanners) {<a name="line.469"></a> -<span class="sourceLineNo">470</span> boolean isFile = kvs.isFileScanner();<a name="line.470"></a> -<span class="sourceLineNo">471</span> if ((!isFile && filesOnly) || (isFile && memOnly)) {<a name="line.471"></a> -<span class="sourceLineNo">472</span> continue;<a name="line.472"></a> -<span class="sourceLineNo">473</span> }<a name="line.473"></a> -<span class="sourceLineNo">474</span><a name="line.474"></a> -<span class="sourceLineNo">475</span> if (kvs.shouldUseScanner(scan, store, expiredTimestampCutoff)) {<a name="line.475"></a> -<span class="sourceLineNo">476</span> scanners.add(kvs);<a name="line.476"></a> -<span class="sourceLineNo">477</span> } else {<a name="line.477"></a> -<span class="sourceLineNo">478</span> kvs.close();<a name="line.478"></a> -<span class="sourceLineNo">479</span> }<a name="line.479"></a> -<span class="sourceLineNo">480</span> }<a name="line.480"></a> -<span class="sourceLineNo">481</span> return scanners;<a name="line.481"></a> -<span class="sourceLineNo">482</span> }<a name="line.482"></a> -<span class="sourceLineNo">483</span><a name="line.483"></a> -<span class="sourceLineNo">484</span> @Override<a name="line.484"></a> -<span class="sourceLineNo">485</span> public Cell peek() {<a name="line.485"></a> -<span class="sourceLineNo">486</span> return heap != null ? heap.peek() : null;<a name="line.486"></a> -<span class="sourceLineNo">487</span> }<a name="line.487"></a> -<span class="sourceLineNo">488</span><a name="line.488"></a> -<span class="sourceLineNo">489</span> @Override<a name="line.489"></a> -<span class="sourceLineNo">490</span> public KeyValue next() {<a name="line.490"></a> -<span class="sourceLineNo">491</span> // throw runtime exception perhaps?<a name="line.491"></a> -<span class="sourceLineNo">492</span> throw new RuntimeException("Never call StoreScanner.next()");<a name="line.492"></a> -<span class="sourceLineNo">493</span> }<a name="line.493"></a> -<span class="sourceLineNo">494</span><a name="line.494"></a> -<span class="sourceLineNo">495</span> @Override<a name="line.495"></a> -<span class="sourceLineNo">496</span> public void close() {<a name="line.496"></a> -<span class="sourceLineNo">497</span> close(true);<a name="line.497"></a> -<span class="sourceLineNo">498</span> }<a name="line.498"></a> -<span class="sourceLineNo">499</span><a name="line.499"></a> -<span class="sourceLineNo">500</span> private void close(boolean withDelayedScannersClose) {<a name="line.500"></a> -<span class="sourceLineNo">501</span> if (this.closing) {<a name="line.501"></a> -<span class="sourceLineNo">502</span> return;<a name="line.502"></a> -<span class="sourceLineNo">503</span> }<a name="line.503"></a> -<span class="sourceLineNo">504</span> if (withDelayedScannersClose) {<a name="line.504"></a> -<span class="sourceLineNo">505</span> this.closing = true;<a name="line.505"></a> -<span class="sourceLineNo">506</span> }<a name="line.506"></a> -<span class="sourceLineNo">507</span> // Under test, we dont have a this.store<a name="line.507"></a> -<span class="sourceLineNo">508</span> if (this.store != null) {<a name="line.508"></a> -<span class="sourceLineNo">509</span> this.store.deleteChangedReaderObserver(this);<a name="line.509"></a> -<span class="sourceLineNo">510</span> }<a name="line.510"></a> -<span class="sourceLineNo">511</span> if (withDelayedScannersClose) {<a name="line.511"></a> -<span class="sourceLineNo">512</span> clearAndClose(scannersForDelayedClose);<a name="line.512"></a> -<span class="sourceLineNo">513</span> clearAndClose(memStoreScannersAfterFlush);<a name="line.513"></a> -<span class="sourceLineNo">514</span> if (this.heap != null) {<a name="line.514"></a> -<span class="sourceLineNo">515</span> this.heap.close();<a name="line.515"></a> -<span class="sourceLineNo">516</span> this.currentScanners.clear();<a name="line.516"></a> -<span class="sourceLineNo">517</span> this.heap = null; // CLOSED!<a name="line.517"></a> -<span class="sourceLineNo">518</span> }<a name="line.518"></a> -<span class="sourceLineNo">519</span> } else {<a name="line.519"></a> -<span class="sourceLineNo">520</span> if (this.heap != null) {<a name="line.520"></a> -<span class="sourceLineNo">521</span> this.scannersForDelayedClose.add(this.heap);<a name="line.521"></a> -<span class="sourceLineNo">522</span> this.currentScanners.clear();<a name="line.522"></a> -<span class="sourceLineNo">523</span> this.heap = null;<a name="line.523"></a> -<span class="sourceLineNo">524</span> }<a name="line.524"></a> -<span class="sourceLineNo">525</span> }<a name="line.525"></a> -<span class="sourceLineNo">526</span> }<a name="line.526"></a> -<span class="sourceLineNo">527</span><a name="line.527"></a> -<span class="sourceLineNo">528</span> @Override<a name="line.528"></a> -<span class="sourceLineNo">529</span> public boolean seek(Cell key) throws IOException {<a name="line.529"></a> -<span class="sourceLineNo">530</span> if (checkFlushed()) {<a name="line.530"></a> -<span class="sourceLineNo">531</span> reopenAfterFlush();<a name="line.531"></a> -<span class="sourceLineNo">532</span> }<a name="line.532"></a> -<span class="sourceLineNo">533</span> return this.heap.seek(key);<a name="line.533"></a> -<span class="sourceLineNo">534</span> }<a name="line.534"></a> -<span class="sourceLineNo">535</span><a name="line.535"></a> -<span class="sourceLineNo">536</span> @Override<a name="line.536"></a> -<span class="sourceLineNo">537</span> public boolean next(List<Cell> outResult) throws IOException {<a name="line.537"></a> -<span class="sourceLineNo">538</span> return next(outResult, NoLimitScannerContext.getInstance());<a name="line.538"></a> -<span class="sourceLineNo">539</span> }<a name="line.539"></a> -<span class="sourceLineNo">540</span><a name="line.540"></a> -<span class="sourceLineNo">541</span> /**<a name="line.541"></a> -<span class="sourceLineNo">542</span> * Get the next row of values from this Store.<a name="line.542"></a> -<span class="sourceLineNo">543</span> * @param outResult<a name="line.543"></a> -<span class="sourceLineNo">544</span> * @param scannerContext<a name="line.544"></a> -<span class="sourceLineNo">545</span> * @return true if there are more rows, false if scanner is done<a name="line.545"></a> -<span class="sourceLineNo">546</span> */<a name="line.546"></a> -<span class="sourceLineNo">547</span> @Override<a name="line.547"></a> -<span class="sourceLineNo">548</span> public boolean next(List<Cell> outResult, ScannerContext scannerContext) throws IOException {<a name="line.548"></a> -<span class="sourceLineNo">549</span> if (scannerContext == null) {<a name="line.549"></a> -<span class="sourceLineNo">550</span> throw new IllegalArgumentException("Scanner context cannot be null");<a name="line.550"></a> -<span class="sourceLineNo">551</span> }<a name="line.551"></a> -<span class="sourceLineNo">552</span> if (checkFlushed() && reopenAfterFlush()) {<a name="line.552"></a> -<span class="sourceLineNo">553</span> return scannerContext.setScannerState(NextState.MORE_VALUES).hasMoreValues();<a name="line.553"></a> -<span class="sourceLineNo">554</span> }<a name="line.554"></a> -<span class="sourceLineNo">555</span><a name="line.555"></a> -<span class="sourceLineNo">556</span> // if the heap was left null, then the scanners had previously run out anyways, close and<a name="line.556"></a> -<span class="sourceLineNo">557</span> // return.<a name="line.557"></a> -<span class="sourceLineNo">558</span> if (this.heap == null) {<a name="line.558"></a> -<span class="sourceLineNo">559</span> // By this time partial close should happened because already heap is null<a name="line.559"></a> -<span class="sourceLineNo">560</span> close(false);// Do all cleanup except heap.close()<a name="line.560"></a> -<span class="sourceLineNo">561</span> return scannerContext.setScannerState(NextState.NO_MORE_VALUES).hasMoreValues();<a name="line.561"></a> -<span class="sourceLineNo">562</span> }<a name="line.562"></a> -<span class="sourceLineNo">563</span><a name="line.563"></a> -<span class="sourceLineNo">564</span> Cell cell = this.heap.peek();<a name="line.564"></a> -<span class="sourceLineNo">565</span> if (cell == null) {<a name="line.565"></a> -<span class="sourceLineNo">566</span> close(false);// Do all cleanup except heap.close()<a name="line.566"></a> -<span class="sourceLineNo">567</span> return scannerContext.setScannerState(NextState.NO_MORE_VALUES).hasMoreValues();<a name="line.567"></a> -<span class="sourceLineNo">568</span> }<a name="line.568"></a> -<span class="sourceLineNo">569</span><a name="line.569"></a> -<span class="sourceLineNo">570</span> // only call setRow if the row changes; avoids confusing the query matcher<a name="line.570"></a> -<span class="sourceLineNo">571</span> // if scanning intra-row<a name="line.571"></a> -<span class="sourceLineNo">572</span><a name="line.572"></a> -<span class="sourceLineNo">573</span> // If no limits exists in the scope LimitScope.Between_Cells then we are sure we are changing<a name="line.573"></a> -<span class="sourceLineNo">574</span> // rows. Else it is possible we are still traversing the same row so we must perform the row<a name="line.574"></a> -<span class="sourceLineNo">575</span> // comparison.<a name="line.575"></a> -<span class="sourceLineNo">576</span> if (!scannerContext.hasAnyLimit(LimitScope.BETWEEN_CELLS) || matcher.currentRow() == null) {<a name="line.576"></a> -<span class="sourceLineNo">577</span> this.countPerRow = 0;<a name="line.577"></a> -<span class="sourceLineNo">578</span> matcher.setToNewRow(cell);<a name="line.578"></a> -<span class="sourceLineNo">579</span> }<a name="line.579"></a> -<span class="sourceLineNo">580</span><a name="line.580"></a> -<span class="sourceLineNo">581</span> // Clear progress away unless invoker has indicated it should be kept.<a name="line.581"></a> -<span class="sourceLineNo">582</span> if (!scannerContext.getKeepProgress()) {<a name="line.582"></a> -<span class="sourceLineNo">583</span> scannerContext.clearProgress();<a name="line.583"></a> -<span class="sourceLineNo">584</span> }<a name="line.584"></a> -<span class="sourceLineNo">585</span><a name="line.585"></a> -<span class="sourceLineNo">586</span> // Only do a sanity-check if store and comparator are available.<a name="line.586"></a> -<span class="sourceLineNo">587</span> CellComparator comparator = store != null ? store.getComparator() : null;<a name="line.587"></a> -<span class="sourceLineNo">588</span><a name="line.588"></a> -<span class="sourceLineNo">589</span> int count = 0;<a name="line.589"></a> -<span class="sourceLineNo">590</span> long totalBytesRead = 0;<a name="line.590"></a> -<span class="sourceLineNo">591</span><a name="line.591"></a> -<span class="sourceLineNo">592</span> LOOP: do {<a name="line.592"></a> -<span class="sourceLineNo">593</span> // Update and check the time limit based on the configured value of cellsPerTimeoutCheck<a name="line.593"></a> -<span class="sourceLineNo">594</span> if ((kvsScanned % cellsPerHeartbeatCheck == 0)) {<a name="line.594"></a> -<span class="sourceLineNo">595</span> scannerContext.updateTimeProgress();<a name="line.595"></a> -<span class="sourceLineNo">596</span> if (scannerContext.checkTimeLimit(LimitScope.BETWEEN_CELLS)) {<a name="line.596"></a> -<span class="sourceLineNo">597</span> scannerContext.setPeekedCellInHeartbeat(prevCell);<a name="line.597"></a> -<span class="sourceLineNo">598</span> return scannerContext.setScannerState(NextState.TIME_LIMIT_REACHED).hasMoreValues();<a name="line.598"></a> -<span class="sourceLineNo">599</span> }<a name="line.599"></a> -<span class="sourceLineNo">600</span> }<a name="line.600"></a> -<span class="sourceLineNo">601</span> // Do object compare - we set prevKV from the same heap.<a name="line.601"></a> -<span class="sourceLineNo">602</span> if (prevCell != cell) {<a name="line.602"></a> -<span class="sourceLineNo">603</span> ++kvsScanned;<a name="line.603"></a> -<span class="sourceLineNo">604</span> }<a name="line.604"></a> -<span class="sourceLineNo">605</span> checkScanOrder(prevCell, cell, comparator);<a name="line.605"></a> -<span class="sourceLineNo">606</span> int cellSize = CellUtil.estimatedSerializedSizeOf(cell);<a name="line.606"></a> -<span class="sourceLineNo">607</span> bytesRead += cellSize;<a name="line.607"></a> -<span class="sourceLineNo">608</span> prevCell = cell;<a name="line.608"></a> +<span class="sourceLineNo">143</span> // A flag whether use pread for scan<a name="line.143"></a> +<span class="sourceLineNo">144</span> // it maybe changed if we use Scan.ReadType.DEFAULT and we have read lots of data.<a name="line.144"></a> +<span class="sourceLineNo">145</span> private boolean scanUsePread;<a name="line.145"></a> +<span class="sourceLineNo">146</span> // Indicates whether there was flush during the course of the scan<a name="line.146"></a> +<span class="sourceLineNo">147</span> private volatile boolean flushed = false;<a name="line.147"></a> +<span class="sourceLineNo">148</span> // generally we get one file from a flush<a name="line.148"></a> +<span class="sourceLineNo">149</span> private final List<StoreFile> flushedStoreFiles = new ArrayList<>(1);<a name="line.149"></a> +<span class="sourceLineNo">150</span> // Since CompactingMemstore is now default, we get three memstore scanners from a flush<a name="line.150"></a> +<span class="sourceLineNo">151</span> private final List<KeyValueScanner> memStoreScannersAfterFlush = new ArrayList<>(3);<a name="line.151"></a> +<span class="sourceLineNo">152</span> // The current list of scanners<a name="line.152"></a> +<span class="sourceLineNo">153</span> @VisibleForTesting<a name="line.153"></a> +<span class="sourceLineNo">154</span> final List<KeyValueScanner> currentScanners = new ArrayList<>();<a name="line.154"></a> +<span class="sourceLineNo">155</span> // flush update lock<a name="line.155"></a> +<span class="sourceLineNo">156</span> private final ReentrantLock flushLock = new ReentrantLock();<a name="line.156"></a> +<span class="sourceLineNo">157</span><a name="line.157"></a> +<span class="sourceLineNo">158</span> protected final long readPt;<a name="line.158"></a> +<span class="sourceLineNo">159</span> private boolean topChanged = false;<a name="line.159"></a> +<span class="sourceLineNo">160</span><a name="line.160"></a> +<span class="sourceLineNo">161</span> // used by the injection framework to test race between StoreScanner construction and compaction<a name="line.161"></a> +<span class="sourceLineNo">162</span> enum StoreScannerCompactionRace {<a name="line.162"></a> +<span class="sourceLineNo">163</span> BEFORE_SEEK,<a name="line.163"></a> +<span class="sourceLineNo">164</span> AFTER_SEEK,<a name="line.164"></a> +<span class="sourceLineNo">165</span> COMPACT_COMPLETE<a name="line.165"></a> +<span class="sourceLineNo">166</span> }<a name="line.166"></a> +<span class="sourceLineNo">167</span><a name="line.167"></a> +<span class="sourceLineNo">168</span> /** An internal constructor. */<a name="line.168"></a> +<span class="sourceLineNo">169</span> protected StoreScanner(Store store, Scan scan, final ScanInfo scanInfo,<a name="line.169"></a> +<span class="sourceLineNo">170</span> final NavigableSet<byte[]> columns, long readPt, boolean cacheBlocks, ScanType scanType) {<a name="line.170"></a> +<span class="sourceLineNo">171</span> this.readPt = readPt;<a name="line.171"></a> +<span class="sourceLineNo">172</span> this.store = store;<a name="line.172"></a> +<span class="sourceLineNo">173</span> this.cacheBlocks = cacheBlocks;<a name="line.173"></a> +<span class="sourceLineNo">174</span> get = scan.isGetScan();<a name="line.174"></a> +<span class="sourceLineNo">175</span> int numCol = columns == null ? 0 : columns.size();<a name="line.175"></a> +<span class="sourceLineNo">176</span> explicitColumnQuery = numCol > 0;<a name="line.176"></a> +<span class="sourceLineNo">177</span> this.scan = scan;<a name="line.177"></a> +<span class="sourceLineNo">178</span> this.now = EnvironmentEdgeManager.currentTime();<a name="line.178"></a> +<span class="sourceLineNo">179</span> this.oldestUnexpiredTS = scan.isRaw() ? 0L : now - scanInfo.getTtl();<a name="line.179"></a> +<span class="sourceLineNo">180</span> this.minVersions = scanInfo.getMinVersions();<a name="line.180"></a> +<span class="sourceLineNo">181</span><a name="line.181"></a> +<span class="sourceLineNo">182</span> // We look up row-column Bloom filters for multi-column queries as part of<a name="line.182"></a> +<span class="sourceLineNo">183</span> // the seek operation. However, we also look the row-column Bloom filter<a name="line.183"></a> +<span class="sourceLineNo">184</span> // for multi-row (non-"get") scans because this is not done in<a name="line.184"></a> +<span class="sourceLineNo">185</span> // StoreFile.passesBloomFilter(Scan, SortedSet<byte[]>).<a name="line.185"></a> +<span class="sourceLineNo">186</span> this.useRowColBloom = numCol > 1 || (!get && numCol == 1);<a name="line.186"></a> +<span class="sourceLineNo">187</span><a name="line.187"></a> +<span class="sourceLineNo">188</span> this.maxRowSize = scanInfo.getTableMaxRowSize();<a name="line.188"></a> +<span class="sourceLineNo">189</span> if (get) {<a name="line.189"></a> +<span class="sourceLineNo">190</span> this.readType = Scan.ReadType.PREAD;<a name="line.190"></a> +<span class="sourceLineNo">191</span> this.scanUsePread = true;<a name="line.191"></a> +<span class="sourceLineNo">192</span> } else if(scanType != scanType.USER_SCAN) {<a name="line.192"></a> +<span class="sourceLineNo">193</span> // For compaction scanners never use Pread as already we have stream based scanners on the<a name="line.193"></a> +<span class="sourceLineNo">194</span> // store files to be compacted<a name="line.194"></a> +<span class="sourceLineNo">195</span> this.readType = Scan.ReadType.STREAM;<a name="line.195"></a> +<span class="sourceLineNo">196</span> this.scanUsePread = false;<a name="line.196"></a> +<span class="sourceLineNo">197</span> } else {<a name="line.197"></a> +<span class="sourceLineNo">198</span> if (scan.getReadType() == Scan.ReadType.DEFAULT) {<a name="line.198"></a> +<span class="sourceLineNo">199</span> this.readType = scanInfo.isUsePread() ? Scan.ReadType.PREAD : Scan.ReadType.DEFAULT;<a name="line.199"></a> +<span class="sourceLineNo">200</span> } else {<a name="line.200"></a> +<span class="sourceLineNo">201</span> this.readType = scan.getReadType();<a name="line.201"></a> +<span class="sourceLineNo">202</span> }<a name="line.202"></a> +<span class="sourceLineNo">203</span> // Always start with pread unless user specific stream. Will change to stream later if<a name="line.203"></a> +<span class="sourceLineNo">204</span> // readType is default if the scan keeps running for a long time.<a name="line.204"></a> +<span class="sourceLineNo">205</span> this.scanUsePread = this.readType != Scan.ReadType.STREAM;<a name="line.205"></a> +<span class="sourceLineNo">206</span> }<a name="line.206"></a> +<span class="sourceLineNo">207</span> this.preadMaxBytes = scanInfo.getPreadMaxBytes();<a name="line.207"></a> +<span class="sourceLineNo">208</span> this.cellsPerHeartbeatCheck = scanInfo.getCellsPerTimeoutCheck();<a name="line.208"></a> +<span class="sourceLineNo">209</span> // Parallel seeking is on if the config allows and more there is more than one store file.<a name="line.209"></a> +<span class="sourceLineNo">210</span> if (this.store != null && this.store.getStorefilesCount() > 1) {<a name="line.210"></a> +<span class="sourceLineNo">211</span> RegionServerServices rsService = ((HStore) store).getHRegion().getRegionServerServices();<a name="line.211"></a> +<span class="sourceLineNo">212</span> if (rsService != null && scanInfo.isParallelSeekEnabled()) {<a name="line.212"></a> +<span class="sourceLineNo">213</span> this.parallelSeekEnabled = true;<a name="line.213"></a> +<span class="sourceLineNo">214</span> this.executor = rsService.getExecutorService();<a name="line.214"></a> +<span class="sourceLineNo">215</span> }<a name="line.215"></a> +<span class="sourceLineNo">216</span> }<a name="line.216"></a> +<span class="sourceLineNo">217</span> }<a name="line.217"></a> +<span class="sourceLineNo">218</span><a name="line.218"></a> +<span class="sourceLineNo">219</span> private void addCurrentScanners(List<? extends KeyValueScanner> scanners) {<a name="line.219"></a> +<span class="sourceLineNo">220</span> this.currentScanners.addAll(scanners);<a name="line.220"></a> +<span class="sourceLineNo">221</span> }<a name="line.221"></a> +<span class="sourceLineNo">222</span><a name="line.222"></a> +<span class="sourceLineNo">223</span> /**<a name="line.223"></a> +<span class="sourceLineNo">224</span> * Opens a scanner across memstore, snapshot, and all StoreFiles. Assumes we<a name="line.224"></a> +<span class="sourceLineNo">225</span> * are not in a compaction.<a name="line.225"></a> +<span class="sourceLineNo">226</span> *<a name="line.226"></a> +<span class="sourceLineNo">227</span> * @param store who we scan<a name="line.227"></a> +<span class="sourceLineNo">228</span> * @param scan the spec<a name="line.228"></a> +<span class="sourceLineNo">229</span> * @param columns which columns we are scanning<a name="line.229"></a> +<span class="sourceLineNo">230</span> * @throws IOException<a name="line.230"></a> +<span class="sourceLineNo">231</span> */<a name="line.231"></a> +<span class="sourceLineNo">232</span> public StoreScanner(Store store, ScanInfo scanInfo, Scan scan, final NavigableSet<byte[]> columns,<a name="line.232"></a> +<span class="sourceLineNo">233</span> long readPt)<a name="line.233"></a> +<span class="sourceLineNo">234</span> throws IOException {<a name="line.234"></a> +<span class="sourceLineNo">235</span> this(store, scan, scanInfo, columns, readPt, scan.getCacheBlocks(), ScanType.USER_SCAN);<a name="line.235"></a> +<span class="sourceLineNo">236</span> if (columns != null && scan.isRaw()) {<a name="line.236"></a> +<span class="sourceLineNo">237</span> throw new DoNotRetryIOException("Cannot specify any column for a raw scan");<a name="line.237"></a> +<span class="sourceLineNo">238</span> }<a name="line.238"></a> +<span class="sourceLineNo">239</span> matcher = UserScanQueryMatcher.create(scan, scanInfo, columns, oldestUnexpiredTS, now,<a name="line.239"></a> +<span class="sourceLineNo">240</span> store.getCoprocessorHost());<a name="line.240"></a> +<span class="sourceLineNo">241</span><a name="line.241"></a> +<span class="sourceLineNo">242</span> this.store.addChangedReaderObserver(this);<a name="line.242"></a> +<span class="sourceLineNo">243</span><a name="line.243"></a> +<span class="sourceLineNo">244</span> try {<a name="line.244"></a> +<span class="sourceLineNo">245</span> // Pass columns to try to filter out unnecessary StoreFiles.<a name="line.245"></a> +<span class="sourceLineNo">246</span> List<KeyValueScanner> scanners = getScannersNoCompaction();<a name="line.246"></a> +<span class="sourceLineNo">247</span><a name="line.247"></a> +<span class="sourceLineNo">248</span> // Seek all scanners to the start of the Row (or if the exact matching row<a name="line.248"></a> +<span class="sourceLineNo">249</span> // key does not exist, then to the start of the next matching Row).<a name="line.249"></a> +<span class="sourceLineNo">250</span> // Always check bloom filter to optimize the top row seek for delete<a name="line.250"></a> +<span class="sourceLineNo">251</span> // family marker.<a name="line.251"></a> +<span class="sourceLineNo">252</span> seekScanners(scanners, matcher.getStartKey(), explicitColumnQuery && lazySeekEnabledGlobally,<a name="line.252"></a> +<span class="sourceLineNo">253</span> parallelSeekEnabled);<a name="line.253"></a> +<span class="sourceLineNo">254</span><a name="line.254"></a> +<span class="sourceLineNo">255</span> // set storeLimit<a name="line.255"></a> +<span class="sourceLineNo">256</span> this.storeLimit = scan.getMaxResultsPerColumnFamily();<a name="line.256"></a> +<span class="sourceLineNo">257</span><a name="line.257"></a> +<span class="sourceLineNo">258</span> // set rowOffset<a name="line.258"></a> +<span class="sourceLineNo">259</span> this.storeOffset = scan.getRowOffsetPerColumnFamily();<a name="line.259"></a> +<span class="sourceLineNo">260</span> addCurrentScanners(scanners);<a name="line.260"></a> +<span class="sourceLineNo">261</span> // Combine all seeked scanners with a heap<a name="line.261"></a> +<span class="sourceLineNo">262</span> resetKVHeap(scanners, store.getComparator());<a name="line.262"></a> +<span class="sourceLineNo">263</span> } catch (IOException e) {<a name="line.263"></a> +<span class="sourceLineNo">264</span> // remove us from the HStore#changedReaderObservers here or we'll have no chance to<a name="line.264"></a> +<span class="sourceLineNo">265</span> // and might cause memory leak<a name="line.265"></a> +<span class="sourceLineNo">266</span> this.store.deleteChangedReaderObserver(this);<a name="line.266"></a> +<span class="sourceLineNo">267</span> throw e;<a name="line.267"></a> +<span class="sourceLineNo">268</span> }<a name="line.268"></a> +<span class="sourceLineNo">269</span> }<a name="line.269"></a> +<span class="sourceLineNo">270</span><a name="line.270"></a> +<span class="sourceLineNo">271</span> /**<a name="line.271"></a> +<span class="sourceLineNo">272</span> * Used for compactions.<p><a name="line.272"></a> +<span class="sourceLineNo">273</span> *<a name="line.273"></a> +<span class="sourceLineNo">274</span> * Opens a scanner across specified StoreFiles.<a name="line.274"></a> +<span class="sourceLineNo">275</span> * @param store who we scan<a name="line.275"></a> +<span class="sourceLineNo">276</span> * @param scan the spec<a name="line.276"></a> +<span class="sourceLineNo">277</span> * @param scanners ancillary scanners<a name="line.277"></a> +<span class="sourceLineNo">278</span> * @param smallestReadPoint the readPoint that we should use for tracking<a name="line.278"></a> +<span class="sourceLineNo">279</span> * versions<a name="line.279"></a> +<span class="sourceLineNo">280</span> */<a name="line.280"></a> +<span class="sourceLineNo">281</span> public StoreScanner(Store store, ScanInfo scanInfo, Scan scan,<a name="line.281"></a> +<span class="sourceLineNo">282</span> List<? extends KeyValueScanner> scanners, ScanType scanType,<a name="line.282"></a> +<span class="sourceLineNo">283</span> long smallestReadPoint, long earliestPutTs) throws IOException {<a name="line.283"></a> +<span class="sourceLineNo">284</span> this(store, scanInfo, scan, scanners, scanType, smallestReadPoint, earliestPutTs, null, null);<a name="line.284"></a> +<span class="sourceLineNo">285</span> }<a name="line.285"></a> +<span class="sourceLineNo">286</span><a name="line.286"></a> +<span class="sourceLineNo">287</span> /**<a name="line.287"></a> +<span class="sourceLineNo">288</span> * Used for compactions that drop deletes from a limited range of rows.<p><a name="line.288"></a> +<span class="sourceLineNo">289</span> *<a name="line.289"></a> +<span class="sourceLineNo">290</span> * Opens a scanner across specified StoreFiles.<a name="line.290"></a> +<span class="sourceLineNo">291</span> * @param store who we scan<a name="line.291"></a> +<span class="sourceLineNo">292</span> * @param scan the spec<a name="line.292"></a> +<span class="sourceLineNo">293</span> * @param scanners ancillary scanners<a name="line.293"></a> +<span class="sourceLineNo">294</span> * @param smallestReadPoint the readPoint that we should use for tracking versions<a name="line.294"></a> +<span class="sourceLineNo">295</span> * @param dropDeletesFromRow The inclusive left bound of the range; can be EMPTY_START_ROW.<a name="line.295"></a> +<span class="sourceLineNo">296</span> * @param dropDeletesToRow The exclusive right bound of the range; can be EMPTY_END_ROW.<a name="line.296"></a> +<span class="sourceLineNo">297</span> */<a name="line.297"></a> +<span class="sourceLineNo">298</span> public StoreScanner(Store store, ScanInfo scanInfo, Scan scan,<a name="line.298"></a> +<span class="sourceLineNo">299</span> List<? extends KeyValueScanner> scanners, long smallestReadPoint, long earliestPutTs,<a name="line.299"></a> +<span class="sourceLineNo">300</span> byte[] dropDeletesFromRow, byte[] dropDeletesToRow) throws IOException {<a name="line.300"></a> +<span class="sourceLineNo">301</span> this(store, scanInfo, scan, scanners, ScanType.COMPACT_RETAIN_DELETES, smallestReadPoint,<a name="line.301"></a> +<span class="sourceLineNo">302</span> earliestPutTs, dropDeletesFromRow, dropDeletesToRow);<a name="line.302"></a> +<span class="sourceLineNo">303</span> }<a name="line.303"></a> +<span class="sourceLineNo">304</span><a name="line.304"></a> +<span class="sourceLineNo">305</span> private StoreScanner(Store store, ScanInfo scanInfo, Scan scan,<a name="line.305"></a> +<span class="sourceLineNo">306</span> List<? extends KeyValueScanner> scanners, ScanType scanT
<TRUNCATED>
