http://git-wip-us.apache.org/repos/asf/hbase-site/blob/47be635d/devapidocs/src-html/org/apache/hadoop/hbase/regionserver/StoreScanner.html ---------------------------------------------------------------------- diff --git a/devapidocs/src-html/org/apache/hadoop/hbase/regionserver/StoreScanner.html b/devapidocs/src-html/org/apache/hadoop/hbase/regionserver/StoreScanner.html index ac1ff82..bc8bc72 100644 --- a/devapidocs/src-html/org/apache/hadoop/hbase/regionserver/StoreScanner.html +++ b/devapidocs/src-html/org/apache/hadoop/hbase/regionserver/StoreScanner.html @@ -56,1127 +56,1126 @@ <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 org.apache.hadoop.hbase.shaded.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>/**<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> // 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 scanType, long smallestReadPoint,<a name="line.306"></a> -<span class="sourceLineNo">307</span> long earliestPutTs, byte[] dropDeletesFromRow, byte[] dropDeletesToRow) throws IOException {<a name="line.307"></a> -<span class="sourceLineNo">308</span> this(store, scan, scanInfo, null,<a name="line.308"></a> -<span class="sourceLineNo">309</span> ((HStore) store).getHRegion().getReadPoint(IsolationLevel.READ_COMMITTED), false, scanType);<a name="line.309"></a> -<span class="sourceLineNo">310</span> if (scan.hasFilter() || (scan.getStartRow() != null && scan.getStartRow().length > 0)<a name="line.310"></a> -<span class="sourceLineNo">311</span> || (scan.getStopRow() != null && scan.getStopRow().length > 0)<a name="line.311"></a> -<span class="sourceLineNo">312</span> || !scan.getTimeRange().isAllTime()) {<a name="line.312"></a> -<span class="sourceLineNo">313</span> // use legacy query matcher since we do not consider the scan object in our code. Only used to<a name="line.313"></a> -<span class="sourceLineNo">314</span> // keep compatibility for coprocessor.<a name="line.314"></a> -<span class="sourceLineNo">315</span> matcher = LegacyScanQueryMatcher.create(scan, scanInfo, null, scanType, smallestReadPoint,<a name="line.315"></a> -<span class="sourceLineNo">316</span> earliestPutTs, oldestUnexpiredTS, now, dropDeletesFromRow, dropDeletesToRow,<a name="line.316"></a> -<span class="sourceLineNo">317</span> store.getCoprocessorHost());<a name="line.317"></a> -<span class="sourceLineNo">318</span> } else {<a name="line.318"></a> -<span class="sourceLineNo">319</span> matcher = CompactionScanQueryMatcher.create(scanInfo, scanType, smallestReadPoint,<a name="line.319"></a> -<span class="sourceLineNo">320</span> earliestPutTs, oldestUnexpiredTS, now, dropDeletesFromRow, dropDeletesToRow,<a name="line.320"></a> -<span class="sourceLineNo">321</span> store.getCoprocessorHost());<a name="line.321"></a> -<span class="sourceLineNo">322</span> }<a name="line.322"></a> -<span class="sourceLineNo">323</span><a name="line.323"></a> -<span class="sourceLineNo">324</span> // Filter the list of scanners using Bloom filters, time range, TTL, etc.<a name="line.324"></a> -<span class="sourceLineNo">325</span> scanners = selectScannersFrom(scanners);<a name="line.325"></a> -<span class="sourceLineNo">326</span><a name="line.326"></a> -<span class="sourceLineNo">327</span> // Seek all scanners to the initial key<a name="line.327"></a> -<span class="sourceLineNo">328</span> seekScanners(scanners, matcher.getStartKey(), false, parallelSeekEnabled);<a name="line.328"></a> -<span class="sourceLineNo">329</span> addCurrentScanners(scanners);<a name="line.329"></a> -<span class="sourceLineNo">330</span> // Combine all seeked scanners with a heap<a name="line.330"></a> -<span class="sourceLineNo">331</span> resetKVHeap(scanners, store.getComparator());<a name="line.331"></a> -<span class="sourceLineNo">332</span> }<a name="line.332"></a> -<span class="sourceLineNo">333</span><a name="line.333"></a> -<span class="sourceLineNo">334</span> @VisibleForTesting<a name="line.334"></a> -<span class="sourceLineNo">335</span> StoreScanner(final Scan scan, ScanInfo scanInfo,<a name="line.335"></a> -<span class="sourceLineNo">336</span> ScanType scanType, final NavigableSet<byte[]> columns,<a name="line.336"></a> -<span class="sourceLineNo">337</span> final List<? extends KeyValueScanner> scanners) throws IOException {<a name="line.337"></a> -<span class="sourceLineNo">338</span> this(scan, scanInfo, scanType, columns, scanners,<a name="line.338"></a> -<span class="sourceLineNo">339</span> HConstants.LATEST_TIMESTAMP,<a name="line.339"></a> -<span class="sourceLineNo">340</span> // 0 is passed as readpoint because the test bypasses Store<a name="line.340"></a> -<span class="sourceLineNo">341</span> 0);<a name="line.341"></a> -<span class="sourceLineNo">342</span> }<a name="line.342"></a> -<span class="sourceLineNo">343</span><a name="line.343"></a> -<span class="sourceLineNo">344</span> @VisibleForTesting<a name="line.344"></a> -<span class="sourceLineNo">345</span> StoreScanner(final Scan scan, ScanInfo scanInfo,<a name="line.345"></a> -<span class="sourceLineNo">346</span> ScanType scanType, final NavigableSet<byte[]> columns,<a name="line.346"></a> -<span class="sourceLineNo">347</span> final List<? extends KeyValueScanner> scanners, long earliestPutTs)<a name="line.347"></a> -<span class="sourceLineNo">348</span> throws IOException {<a name="line.348"></a> -<span class="sourceLineNo">349</span> this(scan, scanInfo, scanType, columns, scanners, earliestPutTs,<a name="line.349"></a> -<span class="sourceLineNo">350</span> // 0 is passed as readpoint because the test bypasses Store<a name="line.350"></a> -<span class="sourceLineNo">351</span> 0);<a name="line.351"></a> -<span class="sourceLineNo">352</span> }<a name="line.352"></a> -<span class="sourceLineNo">353</span><a name="line.353"></a> -<span class="sourceLineNo">354</span> public StoreScanner(final Scan scan, ScanInfo scanInfo, ScanType scanType,<a name="line.354"></a> -<span class="sourceLineNo">355</span> final NavigableSet<byte[]> columns, final List<? extends KeyValueScanner> scanners, long earliestPutTs,<a name="line.355"></a> -<span class="sourceLineNo">356</span> long readPt) throws IOException {<a name="line.356"></a> -<span class="sourceLineNo">357</span> this(null, scan, scanInfo, columns, readPt,<a name="line.357"></a> -<span class="sourceLineNo">358</span> scanType == ScanType.USER_SCAN ? scan.getCacheBlocks() : false, scanType);<a name="line.358"></a> -<span class="sourceLineNo">359</span> if (scanType == ScanType.USER_SCAN) {<a name="line.359"></a> -<span class="sourceLineNo">360</span> this.matcher = UserScanQueryMatcher.create(scan, scanInfo, columns, oldestUnexpiredTS, now,<a name="line.360"></a> -<span class="sourceLineNo">361</span> null);<a name="line.361"></a> -<span class="sourceLineNo">362</span> } else {<a name="line.362"></a> -<span class="sourceLineNo">363</span> if (scan.hasFilter() || (scan.getStartRow() != null && scan.getStartRow().length > 0)<a name="line.363"></a> -<span class="sourceLineNo">364</span> || (scan.getStopRow() != null && scan.getStopRow().length > 0)<a name="line.364"></a> -<span class="sourceLineNo">365</span> || !scan.getTimeRange().isAllTime() || columns != null) {<a name="line.365"></a> -<span class="sourceLineNo">366</span> // use legacy query matcher since we do not consider the scan object in our code. Only used<a name="line.366"></a> -<span class="sourceLineNo">367</span> // to keep compatibility for coprocessor.<a name="line.367"></a> -<span class="sourceLineNo">368</span> matcher = LegacyScanQueryMatcher.create(scan, scanInfo, columns, scanType, Long.MAX_VALUE,<a name="line.368"></a> -<span class="sourceLineNo">369</span> earliestPutTs, oldestUnexpiredTS, now, null, null, store.getCoprocessorHost());<a name="line.369"></a> -<span class="sourceLineNo">370</span> } else {<a name="line.370"></a> -<span class="sourceLineNo">371</span> this.matcher = CompactionScanQueryMatcher.create(scanInfo, scanType, Long.MAX_VALUE,<a name="line.371"></a> -<span class="sourceLineNo">372</span> earliestPutTs, oldestUnexpiredTS, now, null, null, null);<a name="line.372"></a> -<span class="sourceLineNo">373</span> }<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> // Seek all scanners to the initial key<a name="line.376"></a> -<span class="sourceLineNo">377</span> seekScanners(scanners, matcher.getStartKey(), false, parallelSeekEnabled);<a name="line.377"></a> -<span class="sourceLineNo">378</span> addCurrentScanners(scanners);<a name="line.378"></a> -<span class="sourceLineNo">379</span> resetKVHeap(scanners, scanInfo.getComparator());<a name="line.379"></a> -<span class="sourceLineNo">380</span> }<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> * Get a filtered list of scanners. Assumes we are not in a compaction.<a name="line.383"></a> -<span class="sourceLineNo">384</span> * @return list of scanners to seek<a name="line.384"></a> -<span class="sourceLineNo">385</span> */<a name="line.385"></a> -<span class="sourceLineNo">386</span> private List<KeyValueScanner> getScannersNoCompaction() throws IOException {<a name="line.386"></a> -<span class="sourceLineNo">387</span> return selectScannersFrom(<a name="line.387"></a> -<span class="sourceLineNo">388</span> store.getScanners(cacheBlocks, scanUsePread, false, matcher, scan.getStartRow(),<a name="line.388"></a> -<span class="sourceLineNo">389</span> scan.includeStartRow(), scan.getStopRow(), scan.includeStopRow(), this.readPt));<a name="line.389"></a> -<span class="sourceLineNo">390</span> }<a name="line.390"></a> -<span class="sourceLineNo">391</span><a name="line.391"></a> -<span class="sourceLineNo">392</span> @VisibleForTesting<a name="line.392"></a> -<span class="sourceLineNo">393</span> boolean isScanUsePread() {<a name="line.393"></a> -<span class="sourceLineNo">394</span> return this.scanUsePread;<a name="line.394"></a> -<span class="sourceLineNo">395</span> }<a name="line.395"></a> -<span class="sourceLineNo">396</span> /**<a name="line.396"></a> -<span class="sourceLineNo">397</span> * Seek the specified scanners with the given key<a name="line.397"></a> -<span class="sourceLineNo">398</span> * @param scanners<a name="line.398"></a> -<span class="sourceLineNo">399</span> * @param seekKey<a name="line.399"></a> -<span class="sourceLineNo">400</span> * @param isLazy true if using lazy seek<a name="line.400"></a> -<span class="sourceLineNo">401</span> * @param isParallelSeek true if using parallel seek<a name="line.401"></a> -<span class="sourceLineNo">402</span> * @throws IOException<a name="line.402"></a> -<span class="sourceLineNo">403</span> */<a name="line.403"></a> -<span class="sourceLineNo">404</span> protected void seekScanners(List<? extends KeyValueScanner> scanners,<a name="line.404"></a> -<span class="sourceLineNo">405</span> Cell seekKey, boolean isLazy, boolean isParallelSeek)<a name="line.405"></a> -<span class="sourceLineNo">406</span> throws IOException {<a name="line.406"></a> -<span class="sourceLineNo">407</span> // Seek all scanners to the start of the Row (or if the exact matching row<a name="line.407"></a> -<span class="sourceLineNo">408</span> // key does not exist, then to the start of the next matching Row).<a name="line.408"></a> -<span class="sourceLineNo">409</span> // Always check bloom filter to optimize the top row seek for delete<a name="line.409"></a> -<span class="sourceLineNo">410</span> // family marker.<a name="line.410"></a> -<span class="sourceLineNo">411</span> if (isLazy) {<a name="line.411"></a> -<span class="sourceLineNo">412</span> for (KeyValueScanner scanner : scanners) {<a name="line.412"></a> -<span class="sourceLineNo">413</span> scanner.requestSeek(seekKey, false, true);<a name="line.413"></a> -<span class="sourceLineNo">414</span> }<a name="line.414"></a> -<span class="sourceLineNo">415</span> } else {<a name="line.415"></a> -<span class="sourceLineNo">416</span> if (!isParallelSeek) {<a name="line.416"></a> -<span class="sourceLineNo">417</span> long totalScannersSoughtBytes = 0;<a name="line.417"></a> -<span class="sourceLineNo">418</span> for (KeyValueScanner scanner : scanners) {<a name="line.418"></a> -<span class="sourceLineNo">419</span> if (matcher.isUserScan() && totalScannersSoughtBytes >= maxRowSize) {<a name="line.419"></a> -<span class="sourceLineNo">420</span> throw new RowTooBigException("Max row size allowed: " + maxRowSize<a name="line.420"></a> -<span class="sourceLineNo">421</span> + ", but row is bigger than that");<a name="line.421"></a> -<span class="sourceLineNo">422</span> }<a name="line.422"></a> -<span class="sourceLineNo">423</span> scanner.seek(seekKey);<a name="line.423"></a> -<span class="sourceLineNo">424</span> Cell c = scanner.peek();<a name="line.424"></a> -<span class="sourceLineNo">425</span> if (c != null) {<a name="line.425"></a> -<span class="sourceLineNo">426</span> totalScannersSoughtBytes += CellUtil.estimatedSerializedSizeOf(c);<a name="line.426"></a> -<span class="sourceLineNo">427</span> }<a name="line.427"></a> -<span class="sourceLineNo">428</span> }<a name="line.428"></a> -<span class="sourceLineNo">429</span> } else {<a name="line.429"></a> -<span class="sourceLineNo">430</span> parallelSeek(scanners, seekKey);<a name="line.430"></a> -<span class="sourceLineNo">431</span> }<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> protected void resetKVHeap(List<? extends KeyValueScanner> scanners,<a name="line.435"></a> -<span class="sourceLineNo">436</span> CellComparator comparator) throws IOException {<a name="line.436"></a> -<span class="sourceLineNo">437</span> // Combine all seeked scanners with a heap<a name="line.437"></a> -<span class="sourceLineNo">438</span> heap = new KeyValueHeap(scanners, comparator);<a name="line.438"></a> -<span class="sourceLineNo">439</span> }<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> * Filters the given list of scanners using Bloom filter, time range, and<a name="line.442"></a> -<span class="sourceLineNo">443</span> * TTL.<a name="line.443"></a> -<span class="sourceLineNo">444</span> * <p><a name="line.444"></a> -<span class="sourceLineNo">445</span> * Will be overridden by testcase so declared as protected.<a name="line.445"></a> -<span class="sourceLineNo">446</span> */<a name="line.446"></a> -<span class="sourceLineNo">447</span> @VisibleForTesting<a name="line.447"></a> -<span class="sourceLineNo">448</span> protected List<KeyValueScanner> selectScannersFrom(<a name="line.448"></a> -<span class="sourceLineNo">449</span> final List<? extends KeyValueScanner> allScanners) {<a name="line.449"></a> -<span class="sourceLineNo">450</span> boolean memOnly;<a name="line.450"></a> -<span class="sourceLineNo">451</span> boolean filesOnly;<a name="line.451"></a> -<span class="sourceLineNo">452</span> if (scan instanceof InternalScan) {<a name="line.452"></a> -<span class="sourceLineNo">453</span> InternalScan iscan = (InternalScan)scan;<a name="line.453"></a> -<span class="sourceLineNo">454</span> memOnly = iscan.isCheckOnlyMemStore();<a name="line.454"></a> -<span class="sourceLineNo">455</span> filesOnly = iscan.isCheckOnlyStoreFiles();<a name="line.455"></a> -<span class="sourceLineNo">456</span> } else {<a name="line.456"></a> -<span class="sourceLineNo">457</span> memOnly = false;<a name="line.457"></a> -<span class="sourceLineNo">458</span> filesOnly = false;<a name="line.458"></a> -<span class="sourceLineNo">459</span> }<a name="line.459"></a> -<span class="sourceLineNo">460</span><a name="line.460"></a> -<span class="sourceLineNo">461</span> List<KeyValueScanner> scanners = new ArrayList<>(allScanners.size());<a name="line.461"></a> -<span class="sourceLineNo">462</span><a name="line.462"></a> -<span class="sourceLineNo">463</span> // We can only exclude store files based on TTL if minVersions is set to 0.<a name="line.463"></a> -<span class="sourceLineNo">464</span> // Otherwise, we might have to return KVs that have technically expired.<a name="line.464"></a> -<span class="sourceLineNo">465</span> long expiredTimestampCutoff = minVersions == 0 ? oldestUnexpiredTS: Long.MIN_VALUE;<a name="line.465"></a> -<span class="sourceLineNo">466</span><a name="line.466"></a> -<span class="sourceLineNo">467</span> // include only those scan files which pass all filters<a name="line.467"></a> -<span class="sourceLineNo">468</span> for (KeyValueScanner kvs : allScanners) {<a name="line.468"></a> -<span class="sourceLineNo">469</span> boolean isFile = kvs.isFileScanner();<a name="line.469"></a> -<span class="sourceLineNo">470</span> if ((!isFile && filesOnly) || (isFile && memOnly)) {<a name="line.470"></a> -<span class="sourceLineNo">471</span> continue;<a name="line.471"></a> -<span class="sourceLineNo">472</span> }<a name="line.472"></a> -<span class="sourceLineNo">473</span><a name="line.473"></a> -<span class="sourceLineNo">474</span> if (kvs.shouldUseScanner(scan, store, expiredTimestampCutoff)) {<a name="line.474"></a> -<span class="sourceLineNo">475</span> scanners.add(kvs);<a name="line.475"></a> -<span class="sourceLineNo">476</span> } else {<a name="line.476"></a> -<span class="sourceLineNo">477</span> kvs.close();<a name="line.477"></a> -<span class="sourceLineNo">478</span> }<a name="line.478"></a> -<span class="sourceLineNo">479</span> }<a name="line.479"></a> -<span class="sourceLineNo">480</span> return scanners;<a name="line.480"></a> -<span class="sourceLineNo">481</span> }<a name="line.481"></a> -<span class="sourceLineNo">482</span><a name="line.482"></a> -<span class="sourceLineNo">483</span> @Override<a name="line.483"></a> -<span class="sourceLineNo">484</span> public Cell peek() {<a name="line.484"></a> -<span class="sourceLineNo">485</span> return heap != null ? heap.peek() : null;<a name="line.485"></a> -<span class="sourceLineNo">486</span> }<a name="line.486"></a> -<span class="sourceLineNo">487</span><a name="line.487"></a> -<span class="sourceLineNo">488</span> @Override<a name="line.488"></a> -<span class="sourceLineNo">489</span> public KeyValue next() {<a name="line.489"></a> -<span class="sourceLineNo">490</span> // throw runtime exception perhaps?<a name="line.490"></a> -<span class="sourceLineNo">491</span> throw new RuntimeException("Never call StoreScanner.next()");<a name="line.491"></a> -<span class="sourceLineNo">492</span> }<a name="line.492"></a> -<span class="sourceLineNo">493</span><a name="line.493"></a> -<span class="sourceLineNo">494</span> @Override<a name="line.494"></a> -<span class="sourceLineNo">495</span> public void close() {<a name="line.495"></a> -<span class="sourceLineNo">496</span> close(true);<a name="line.496"></a> -<span class="sourceLineNo">497</span> }<a name="line.497"></a> -<span class="sourceLineNo">498</span><a name="line.498"></a> -<span class="sourceLineNo">499</span> private void close(boolean withDelayedScannersClose) {<a name="line.499"></a> -<span class="sourceLineNo">500</span> if (this.closing) {<a name="line.500"></a> -<span class="sourceLineNo">501</span> return;<a name="line.501"></a> -<span class="sourceLineNo">502</span> }<a name="line.502"></a> -<span class="sourceLineNo">503</span> if (withDelayedScannersClose) {<a name="line.503"></a> -<span class="sourceLineNo">504</span> this.closing = true;<a name="line.504"></a> -<span class="sourceLineNo">505</span> }<a name="line.505"></a> -<span class="sourceLineNo">506</span> // Under test, we dont have a this.store<a name="line.506"></a> -<span class="sourceLineNo">507</span> if (this.store != null) {<a name="line.507"></a> -<span class="sourceLineNo">508</span> this.store.deleteChangedReaderObserver(this);<a name="line.508"></a> -<span class="sourceLineNo">509</span> }<a name="line.509"></a> -<span class="sourceLineNo">510</span> if (withDelayedScannersClose) {<a name="line.510"></a> -<span class="sourceLineNo">511</span> clearAndClose(scannersForDelayedClose);<a name="line.511"></a> -<span class="sourceLineNo">512</span> clearAndClose(memStoreScannersAfterFlush);<a name="line.512"></a> -<span class="sourceLineNo">513</span> if (this.heap != null) {<a name="line.513"></a> -<span class="sourceLineNo">514</span> this.heap.close();<a name="line.514"></a> -<span class="sourceLineNo">515</span> this.currentScanners.clear();<a name="line.515"></a> -<span class="sourceLineNo">516</span> this.heap = null; // CLOSED!<a name="line.516"></a> -<span class="sourceLineNo">517</span> }<a name="line.517"></a> -<span class="sourceLineNo">518</span> } else {<a name="line.518"></a> -<span class="sourceLineNo">519</span> if (this.heap != null) {<a name="line.519"></a> -<span class="sourceLineNo">520</span> this.scannersForDelayedClose.add(this.heap);<a name="line.520"></a> -<span class="sourceLineNo">521</span> this.currentScanners.clear();<a name="line.521"></a> -<span class="sourceLineNo">522</span> this.heap = null;<a name="line.522"></a> -<span class="sourceLineNo">523</span> }<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> @Override<a name="line.527"></a> -<span class="sourceLineNo">528</span> public boolean seek(Cell key) throws IOException {<a name="line.528"></a> -<span class="sourceLineNo">529</span> if (checkFlushed()) {<a name="line.529"></a> -<span class="sourceLineNo">530</span> reopenAfterFlush();<a name="line.530"></a> -<span class="sourceLineNo">531</span> }<a name="line.531"></a> -<span class="sourceLineNo">532</span> return this.heap.seek(key);<a name="line.532"></a> -<span class="sourceLineNo">533</span> }<a name="line.533"></a> -<span class="sourceLineNo">534</span><a name="line.534"></a> -<span class="sourceLineNo">535</span> @Override<a name="line.535"></a> -<span class="sourceLineNo">536</span> public boolean next(List<Cell> outResult) throws IOException {<a name="line.536"></a> -<span class="sourceLineNo">537</span> return next(outResult, NoLimitScannerContext.getInstance());<a name="line.537"></a> -<span class="sourceLineNo">538</span> }<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> * Get the next row of values from this Store.<a name="line.541"></a> -<span class="sourceLineNo">542</span> * @param outResult<a name="line.542"></a> -<span class="sourceLineNo">543</span> * @param scannerContext<a name="line.543"></a> -<span class="sourceLineNo">544</span> * @return true if there are more rows, false if scanner is done<a name="line.544"></a> -<span class="sourceLineNo">545</span> */<a name="line.545"></a> -<span class="sourceLineNo">546</span> @Override<a name="line.546"></a> -<span class="sourceLineNo">547</span> public boolean next(List<Cell> outResult, ScannerContext scannerContext) throws IOException {<a name="line.547"></a> -<span class="sourceLineNo">548</span> if (scannerContext == null) {<a name="line.548"></a> -<span class="sourceLineNo">549</span> throw new IllegalArgumentException("Scanner context cannot be null");<a name="line.549"></a> -<span class="sourceLineNo">550</span> }<a name="line.550"></a> -<span class="sourceLineNo">551</span> if (checkFlushed() && reopenAfterFlush()) {<a name="line.551"></a> -<span class="sourceLineNo">552</span> return scannerContext.setScannerState(NextState.MORE_VALUES).hasMoreValues();<a name="line.552"></a> -<span class="sourceLineNo">553</span> }<a name="line.553"></a> -<span class="sourceLineNo">554</span><a name="line.554"></a> -<span class="sourceLineNo">555</span> // if the heap was left null, then the scanners had previously run out anyways, close and<a name="line.555"></a> -<span class="sourceLineNo">556</span> // return.<a name="line.556"></a> -<span class="sourceLineNo">557</span> if (this.heap == null) {<a name="line.557"></a> -<span class="sourceLineNo">558</span> // By this time partial close should happened because already heap is null<a name="line.558"></a> -<span class="sourceLineNo">559</span> close(false);// Do all cleanup except heap.close()<a name="line.559"></a> -<span class="sourceLineNo">560</span> return scannerContext.setScannerState(NextState.NO_MORE_VALUES).hasMoreValues();<a name="line.560"></a> -<span class="sourceLineNo">561</span> }<a name="line.561"></a> -<span class="sourceLineNo">562</span><a name="line.562"></a> -<span class="sourceLineNo">563</span> Cell cell = this.heap.peek();<a name="line.563"></a> -<span class="sourceLineNo">564</span> if (cell == null) {<a name="line.564"></a> -<span class="sourceLineNo">565</span> close(false);// Do all cleanup except heap.close()<a name="line.565"></a> -<span class="sourceLineNo">566</span> return scannerContext.setScannerState(NextState.NO_MORE_VALUES).hasMoreValues();<a name="line.566"></a> -<span class="sourceLineNo">567</span> }<a name="line.567"></a> -<span class="sourceLineNo">568</span><a name="line.568"></a> -<span class="sourceLineNo">569</span> // only call setRow if the row changes; avoids confusing the query matcher<a name="line.569"></a> -<span class="sourceLineNo">570</span> // if scanning intra-row<a name="line.570"></a> -<span class="sourceLineNo">571</span><a name="line.571"></a> -<span class="sourceLineNo">572</span> // If no limits exists in the scope LimitScope.Between_Cells then we are sure we are changing<a name="line.572"></a> -<span class="sourceLineNo">573</span> // rows. Else it is possible we are still traversing the same row so we must perform the row<a name="line.573"></a> -<span class="sourceLineNo">574</span> // comparison.<a name="line.574"></a> -<span class="sourceLineNo">575</span> if (!scannerContext.hasAnyLimit(LimitScope.BETWEEN_CELLS) || matcher.currentRow() == null) {<a name="line.575"></a> -<span class="sourceLineNo">576</span> this.countPerRow = 0;<a name="line.576"></a> -<span class="sourceLineNo">577</span> matcher.setToNewRow(cell);<a name="line.577"></a> -<span class="sourceLineNo">578</span> }<a name="line.578"></a> -<span class="sourceLineNo">579</span><a name="line.579"></a> -<span class="sourceLineNo">580</span> // Clear progress away unless invoker has indicated it should be kept.<a name="line.580"></a> -<span class="sourceLineNo">581</span> if (!scannerContext.getKeepProgress()) {<a name="line.581"></a> -<span class="sourceLineNo">582</span> scannerContext.clearProgress();<a name="line.582"></a> -<span class="sourceLineNo">583</span> }<a name="line.583"></a> -<span class="sourceLineNo">584</span><a name="line.584"></a> -<span class="sourceLineNo">585</span> // Only do a sanity-check if store and comparator are available.<a name="line.585"></a> -<span class="sourceLineNo">586</span> CellComparator comparator = store != null ? store.getComparator() : null;<a name="line.586"></a> -<span class="sourceLineNo">587</span><a name="line.587"></a> -<span class="sourceLineNo">588</span> int count = 0;<a name="line.588"></a> -<span class="sourceLineNo">589</span> long totalBytesRead = 0;<a name="line.589"></a> -<span class="sourceLineNo">590</span><a name="line.590"></a> -<span class="sourceLineNo">591</span> LOOP: do {<a name="line.591"></a> -<span class="sourceLineNo">592</span> // Update and check the time limit based on the configured value of cellsPerTimeoutCheck<a name="line.592"></a> -<span class="sourceLineNo">593</span> if ((kvsScanned % cellsPerHeartbeatCheck == 0)) {<a name="line.593"></a> -<span class="sourceLineNo">594</span> scannerContext.updateTimeProgress();<a name="line.594"></a> -<span class="sourceLineNo">595</span> if (scannerContext.checkTimeLimit(LimitScope.BETWEEN_CELLS)) {<a name="line.595"></a> -<span class="sourceLineNo">596</span> scannerContext.setPeekedCellInHeartbeat(prevCell);<a name="line.596"></a> -<span class="sourceLineNo">597</span> return scannerContext.setScannerState(NextState.TIME_LIMIT_REACHED).hasMoreValues();<a name="line.597"></a> -<span class="sourceLineNo">598</span> }<a name="line.598"></a> -<span class="sourceLineNo">599</span> }<a name="line.599"></a> -<span class="sourceLineNo">600</span> // Do object compare - we set prevKV from the same heap.<a name="line.600"></a> -<span class="sourceLineNo">601</span> if (prevCell != cell) {<a name="line.601"></a> -<span class="sourceLineNo">602</span> ++kvsScanned;<a name="line.602"></a> -<span class="sourceLineNo">603</span> }<a name="line.603"></a> -<span class="sourceLineNo">604</span> checkScanOrder(prevCell, cell, comparator);<a name="line.604"></a> -<span class="sourceLineNo">605</span> int cellSize = CellUtil.estimatedSerializedSizeOf(cell);<a name="line.605"></a> -<span class="sourceLineNo">606</span> bytesRead += cellSize;<a name="line.606"></a> -<span class="sourceLineNo">607</span> prevCell = cell;<a name="line.607"></a> -<span class="sourceLineNo">608</span> topChanged = false;<a name="line.608"></a> -<span class="sourceLineNo">609</span> ScanQueryMatcher.MatchCode qcode = matcher.match(cell);<a name="line.609"></a> -<span class="sourceLineNo">610</span> switch (qcode) {<a name="line.610"></a> -<span class="sourceLineNo">611</span> case INCLUDE:<a name="line.611"></a> -<span class="sourceLineNo">612</span> case INCLUDE_AND_SEEK_NEXT_ROW:<a name="line.612"></a> -<span class="sourceLineNo">613</span> case INCLUDE_AND_SEEK_NEXT_COL:<a name="line.613"></a> -<span class="sourceLineNo">614</span><a name="line.614"></a> -<span class="sourceLineNo">615</span> Filter f = matcher.getFilter();<a name="line.615"></a> -<span class="sourceLineNo">616</span> if (f != null) {<a name="line.616"></a> -<span class="sourceLineNo">617</span> cell = f.transformCell(cell);<a name="line.617"></a> -<span class="sourceLineNo">618</span> }<a name="line.618"></a> -<span class="sourceLineNo">619</span><a name="line.619"></a> -<span class="sourceLineNo">620</span> this.countPerRow++;<a name="line.620"></a> -<span class="sourceLineNo">621</span> if (storeLimit > -1 && this.countPerRow > (storeLimit + storeOffset)) {<a name="line.621"></a> -<span class="sourceLineNo">622</span> // do what SEEK_NEXT_ROW does.<a name="line.622"></a> -<span class="sourceLineNo">623</span> if (!matcher.moreRowsMayExistAfter(cell)) {<a name="line.623"></a> -<span class="sourceLineNo">624</span> close(false);// Do all cleanup except heap.close()<a name="line.624"></a> -<span class="sourceLineNo">625</span> return scannerContext.setScannerState(NextState.NO_MORE_VALUES).hasMoreValues();<a name="line.625"></a> -<span class="sourceLineNo">626</span> }<a name="line.626"></a> -<span class="sourceLineNo">627</span> matcher.clearCurrentRow();<a name="line.627"></a> -<span class="sourceLineNo">628</span> seekToNextRow(cell);<a name="line.628"></a> -<span class="sourceLineNo">629</span> break LOOP;<a name="line.629"></a> -<span class="sourceLineNo">630</span> }<a name="line.630"></a> -<span class="sourceLineNo">631</span><a name="line.631"></a> -<span class="sourceLineNo">632</span> // add to results only if we have skipped #storeOffset kvs<a name="line.632"></a> -<span class="sourceLineNo">633</span> // also update metric accordingly<a name="line.633"></a> -<span class="sourceLineNo">634</span> if (this.countPerRow > storeOffset) {<a name="line.634"></a> -<span class="sourceLineNo">635</span> outResult.add(cell);<a name="line.635"></a> -<span class="sourceLineNo">636</span><a name="line.636"></a> -<span class="sourceLineNo">637</span> // Update local tracking information<a name="line.637"></a> -<span class="sourceLineNo">638</span> count++;<a name="line.638"></a> -<span class="sourceLineNo">639</span> totalBytesRead += cellSize;<a name="line.639"></a> -<span class="sourceLineNo">640</span><a name="line.640"></a> -<span class="sourceLineNo">641</span> // Update the progress of the scanner context<a name="line.641"></a> -<span class="sourceLineNo">642</span> scannerContext.incrementSizeProgress(cellSize, CellUtil.estimatedHeapSizeOf(cell));<a name="line.642"></a> -<span class="sourceLineNo">643</span> scannerContext.incrementBatchProgress(1);<a name="line.643"></a> -<span class="sourceLineNo">644</span><a name="line.644"></a> -<span class="sourceLineNo">645</span> if (matcher.isUserScan() && totalBytesRead > maxRowSize) {<a name="line.645"></a> -<span class="sourceLineNo">646</span> throw new RowTooBigException(<a name="line.646"></a> -<span class="sourceLineNo">647</span> "Max row size allowed: " + maxRowSize + ", but the row is bigger than that.");<a name="line.647"></a> -<span class="sourceLineNo">648</span> }<a name="line.648"></a> -<span class="sourceLineNo">649</span> }<a name="line.649"></a> -<span class="sourceLineNo">650</span><a name="line.650"></a> -<span class="sourceLineNo">651</span> if (qcode == ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_ROW) {<a name="line.651"></a> -<span class="sourceLineNo">652</span> if (!matcher.moreRowsMayExistAfter(cell)) {<a name="line.652"></a> -<span class="sourceLineNo">653</span> close(false);// Do all cleanup except heap.close()<a name="line.653"></a> -<span class="sourceLineNo">654</span> return scannerContext.setScannerState(NextState.NO_MORE_VALUES).hasMoreValues();<a name="line.654"></a> -<span class="sourceLineNo">655</span> }<a name="line.655"></a> -<span class="sourceLineNo">656</span> matcher.clearCurrentRow();<a name="line.656"></a> -<span class="sourceLineNo">657</span> seekOrSkipToNextRow(cell);<a name="line.657"></a> -<span class="sourceLineNo">658</span> } else if (qcode == ScanQueryMatcher.MatchCode.INCLUDE_AND_SEEK_NEXT_COL) {<a name="line.658"></a> -<span class="sourceLineNo">659</span> seekOrSkipToNextColumn(cell);<a name="line.659"></a> -<span class="sourceLineNo">660</span> } else {<a name="line.660"></a> -<span class="sourceLineNo">661</span> this.heap.next();<a name="line.661"></a> -<span class="sourceLineNo">662</span> }<a name="line.662"></a> -<span class="sourceLineNo">663</span><a name="line.663"></a> -<span class="sourceLineNo">664</span> if (scannerContext.checkBatchLimit(LimitScope.BETWEEN_CELLS)) {<a name="line.664"></a> -<span class="sourceLineNo">665</span> break LOOP;<a name="line.665"></a> -<span class="sourceLineNo">666</span> }<a name="line.666"></a> -<span class="sourceLineNo">667</span> if (scannerContext.checkSizeLimit(LimitScope.BETWEEN_CELLS)) {<a name="line.667"></a> -<span class="sourceLineNo">668</span> break LOOP;<a name="line.668"></a> -<span class="sourceLineNo">669</span> }<a name="line.669"></a> -<span class="sourceLineNo">670</span> continue;<a name="line.670"></a> -<span class="sourceLineNo">671</span><a name="line.671"></a> -<span class="sourceLineNo">672</span> case DONE:<a name="line.672"></a> -<span class="sourceLineNo">673</span> // Optimization for Gets! If DONE, no more to get on this row, early exit!<a name="line.673"></a> -<span class="sourceLineNo">674</span> if (get) {<a name="line.674"></a> -<span class="sourceLineNo">675</span> // Then no more to this row... exit.<a name="line.675"></a> -<span class="sourceLineNo">676</span> close(false);// Do all cleanup except heap.close()<a name="line.676"></a> -<span class="sourceLineNo">677</span> return scannerContext.setScannerState(NextState.NO_MORE_VALUES).hasMoreValues();<a name="line.677"></a> -<span class="sourceLineNo">678</span> }<a name="line.678"></a> -<span class="sourceLineNo">679</span> matcher.clearCurrentRow();<a name="line.679"></a> -<span class="sourceLineNo">680</span> return scannerContext.setScannerState(NextState.MORE_VALUES).hasMoreValues();<a name="line.680"></a> -<span class="sourceLineNo">681</span><a name="line.681"></a> -<span class="sourceLineNo">682</span> case DONE_SCAN:<a name="line.682"></a> -<span class="sourceLineNo">683</span> close(false);// Do all cleanup except heap.close()<a name="line.683"></a> -<span class="sourceLineNo">684</span> return scannerContext.setScannerState(NextState.NO_MORE_VALUES).hasMoreValues();<a name="line.684"></a> -<span class="sourceLineNo">685</span><a name="line.685"></a> -<span class="sourceLineNo">686</span> case SEEK_NEXT_ROW:<a name="line.686"></a> -<span class="sourceLineNo">687</span> // This is just a relatively simple end of scan fix, to short-cut end<a name="line.687"></a> -<span class="sourceLineNo">688</span> // us if there is an endKey in the scan.<a name="line.688"></a> -<span class="sourceLineNo">689</span> if (!matcher.moreRowsMayExistAfter(cell)) {<a name="line.689"></a> -<span class="sourceLineNo">690</span> close(false);// Do all cleanup except heap.close()<a name="line.690"></a> -<span class="sourceLineNo">691</span> return scannerContext.setScannerState(NextState.NO_MORE_VALUES).hasMoreValues();<a name="line.691"></a> -<span class="sourceLineNo">692</span> }<a name="line.692"></a> -<span class="sourceLineNo">693</span> matcher.clearCurrentRow();<a name="line.693"></a> -<span class="sourceLineNo">694</span> seekOrSkipToNextRow(cell);<a name="line.694"></a> -<span class="sourceLineNo">695</span> NextState stateAfterSeekNextRow = needToReturn(outResult);<a name="line.695"></a> -<span class="sourceLineNo">696</span> if (stateAfterSeekNextRow != null) {<a name="line.696"></a> -<span class="sourceLineNo">697</span> return scannerContext.setScannerState(stateAfterSeekNextRow).hasMoreValues();<a name="line.697"></a> -<span class="sourceLineNo">698</span> }<a name="line.698"></a> -<span class="sourceLineNo">699</span> break;<a name="line.699"></a> -<span class="sourceLineNo">700</span><a name="line.700"></a> -<span class="sourceLineNo">701</span> case SEEK_NEXT_COL:<a name="line.701"></a> -<span class="sourceLineNo">702</span> seekOrSkipToNextColumn(cell);<a name="line.702"></a> -<span class="sourceLineNo">703</span> NextState stateAfterSeekNextColumn = needToReturn(outResult);<a name="line.703"></a> -<span class="sourceLineNo">704</span> if (stateAfterSeekNextColumn != null) {<a name="line.704"></a> -<span class="sourceLineNo">705</span> return scannerContext.setScannerState(stateAfterSeekNextColumn).hasMoreValues();<a name="line.705"></a> -<span class="sourceLineNo">706</span> }<a name="line.706"></a> -<span class="sourceLineNo">707</span> break;<a name="line.707"></a> -<span class="sourceLineNo">708</span><a name="line.708"></a> -<span class="sourceLineNo">709</span> case SKIP:<a name="line.709"></a> -<span class="sourceLineNo">710</span> this.heap.next();<a name="line.710"></a> -<span class="sourceLineNo">711</span> break;<a name="line.711"></a> -<span class="sourceLineNo">712</span><a name="line.712"></a> -<span class="sourceLineNo">713</span> case SEEK_NEXT_USING_HINT:<a name="line.713"></a> -<span class="sourceLineNo">714</span> Cell nextKV = matcher.getNextKeyHint(cell);<a name="line.714"></a> -<span class="sourceLineNo">715</span> if (nextKV != null) {<a name="line.715"></a> -<span class="sourceLineNo">716</span> seekAsDirection(nextKV);<a name="line.716"></a> -<span class="sourceLineNo">717</span> NextState stateAfterSeekByHint = needToReturn(outResult);<a name="line.717"></a> -<span class="sourceLineNo">718</span> if (stateAfterSeekByHint != null) {<a name="line.718"></a> -<span class="sourceLineNo">719</span> return scannerContext.setScannerState(stateAfterSeekByHint).hasMoreValues();<a name="line.719"></a> -<span class="sourceLineNo">720</span> }<a name="line.720"></a> -<span class="sourceLineNo">721</span> } else {<a name="line.721"></a> -<span class="sourceLineNo">722</span> heap.next();<a name="line.722"></a> -<span class="sourceLineNo">723</span> }<a name="line.723"></a> -<span class="sourceLineNo">724</span> break;<a name="line.724"></a> -<span class="sourceLineNo">725</span><a name="line.725"></a> -<span class="sourceLineNo">726</span> default:<a name="line.726"></a> -<span class="sourceLineNo">727</span> throw new RuntimeException("UNEXPECTED");<a name="line.727"></a> -<span class="sourceLineNo">728</span> }<a name="line.728"></a> -<span class="sourceLineNo">729</span> } while ((cell = this.heap.peek()) != null);<a name="line.729"></a> -<span class="sourceLineNo">730</span><a name="line.730"></a> -<span class="sourceLineNo">731</span> if (count > 0) {<a name="line.731"></a> -<span class="sourceLineNo">732</span> return scannerContext.setScannerState(NextState.MORE_VALUES).hasMoreValues();<a name="line.732"></a> -<span class="sourceLineNo">733</span> }<a name="line.733"></a> -<span class="sourceLineNo">734</span><a name="line.734"></a> -<span class="sourceLineNo">735</span> // No more keys<a name="line.735"></a> -<span class="sourceLineNo">736</span> close(false);// Do all cleanup except heap.close()<a name="line.736"></a> -<span class="sourceLineNo">737</span> return scannerContext.setScannerState(NextState.NO_MORE_VALUES).hasMoreValues();<a name="line.737"></a> -<span class="sourceLineNo">738</span> }<a name="line.738"></a> -<span class="sourceLineNo">739</span><a name="line.739"></a> -<span class="sourceLineNo">740</span> /**<a name="line.740"></a> -<span class="sourceLineNo">741</span> * If the top cell won't be flushed into disk, the new top cell may be<a name="line.741"></a> -<span class="sourceLineNo">742</span> * changed after #reopenAfterFlush. Because the older top cell only exist<a name="line.742"></a> -<span class="sourceLineNo">743</span> * in the memstore scanner but the memstore scanner is replaced by hfile<a name="line.743"></a> -<span class="sourceLineNo">744</span> * scanner after #reopenAfterFlush. If the row of top cell is changed,<a name="line.744"></a> -<span class="sourceLineNo">745</span> * we should return the current cells. Otherwise, we may return<a name="line.745"></a> -<span class="sourceLineNo">746</span> * the cells across different rows.<a name="line.746"></a> -<span class="sourceLineNo">747</span> * @param outResult the cells which are visible for user scan<a name="line.747"></a> -<span class="sourceLineNo">748</span> * @return null is the top cell doesn't change. Otherwise, the NextState<a name="line.748"></a> -<span class="sourceLineNo">749</span> * to return<a name="line.749"></a> -<span class="sourceLineNo">750</span> */<a name="line.750"></a> -<span class="sourceLineNo">751</span> private NextState needToReturn(List<Cell> outResult) {<a name="line.751"></a> -<span class="sourceLineNo">752</span> if (!outResult.isEmpty() && topChanged) {<a name="line.752"></a> -<span class="sourceLineNo">753</span> return heap.peek() == null ? NextState.NO_MORE_VALUES : NextState.MORE_VALUES;<a name="line.753"></a> -<span class="sourceLineNo">754</span> }<a name="line.754"></a> -<span class="sourceLineNo">755</span> return null;<a name="line.755"></a> -<span class="sourceLineNo">756</span> }<a name="line.756"></a> -<span class="sourceLineNo">757</span><a name="line.757"></a> -<span class="sourceLineNo">758</span> private void seekOrSkipToNextRow(Cell cell) throws IOException {<a name="line.758"></a> -<span class="sourceLineNo">759</span> // If it is a Get Scan, then we know that we are done with this row; there are no more<a name="line.759"></a> -<span class="sourceLineNo">760</span> // rows beyond the current one: don't try to optimize.<a name="line.760"></a> -<span class="sourceLineNo">761</span> if (!get) {<a name="line.761"></a> -<span class="sourceLineNo">762</span> if (trySkipToNextRow(cell)) {<a name="line.762"></a> -<span class="sourceLineNo">763</span> return;<a name="line.763"></a> -<span class="sourceLineNo">764</span> }<a name="line.764"></a> -<span class="sourceLineNo">765</span> }<a name="line.765"></a> -<span class="sourceLineNo">766</span> seekToNextRow(cell);<a name="line.766"></a> -<span class="sourceLineNo">767</span> }<a name="line.767"></a> -<span class="sourceLineNo">768</span><a name="line.768"></a> -<span class="sourceLineNo">769</span> private void seekOrSkipToNextColumn(Cell cell) throws IOException {<a name="line.769"></a> -<span class="sourceLineNo">770</span> if (!trySkipToNextColumn(cell)) {<a name="line.770"></a> -<span class="sourceLineNo">771</span> seekAsDirection(matcher.getKeyForNextColumn(cell));<a name="line.771"></a> -<span class="sourceLineNo">772</span> }<a name="line.772"></a> -<span class="sourceLineNo">773</span> }<a name="line.773"></a> -<span class="sourceLineNo">774</span><a name="line.774"></a> -<span class="sourceLineNo">775</span> /**<a name="line.775"></a> -<span class="sourceLineNo">776</span> * See if we should actually SEEK or rather just SKIP to the next Cell (see HBASE-13109).<a name="line.776"></a> -<span class="sourceLineNo">777</span> * ScanQueryMatcher may issue SEEK hints, such as seek to next column, next row,<a name="line.777"></a> -<span class="sourceLineNo">778</span> * or seek to an arbitrary seek key. This method decides whether a seek is the most efficient<a name="line.778"></a> -<span class="sourceLineNo">779</span> * _actual_ way to get us to the requested cell (SEEKs are more expensive than SKIP, SKIP,<a name="line.779"></a> -<span class="sourceLineNo">780</span> * SKIP inside the current, loaded block).<a name="line.780"></a> -<span class="sourceLineNo">781</span> * It does this by looking at the next indexed key of the current HFile. This key<a name="line.781"></a> -<span class="sourceLineNo">782</span> * is then compared with the _SEEK_ key, where a SEEK key is an artificial 'last possible key<a name="line.782"></a> -<span class="sourceLineNo">783</span> * on the row' (only in here, we avoid actually creating a SEEK key; in the compare we work with<a name="line.783"></a> -<span class="sourceLineNo">784</span> * the current Cell but compare as though it were a seek key; see down in<a name="line.784"></a> -<span class="sourceLineNo">785</span> * matcher.compareKeyForNextRow, etc). If the compare gets us onto the<a name="line.785"></a> -<span class="sourceLineNo">786</span> * next block we *_SEEK, otherwise we just SKIP to the next requested cell.<a name="line.786"></a> -<span class="sourceLineNo">787</span> *<a name="line.787"></a> -<span class="sourceLineNo">788</span> * <p>Other notes:<a name="line.788"></a> -<span class="sourceLineNo">789</span> * <ul><a name="line.789"></a> -<span class="sourceLineNo">790</span> * <li>Rows can straddle block boundaries</li><a name="line.790"></a> -<span class="sourceLineNo">791</span> * <li>Versions of columns can straddle block boundaries (i.e. column C1 at T1 might be in a<a name="line.791"></a> -<span class="sourceLineNo">792</span> * different block than column C1 at T2)</li><a name="line.792"></a> -<span class="sourceLineNo">793</span> * <li>We want to SKIP if the chance is high that we'll find the desired Cell after a<a name="line.793"></a> -<span class="sourceLineNo">794</span> * few SKIPs...</li><a name="line.794"></a> -<span class="sourceLineNo">795</span> * <li>We want to SEEK when the chance is high that we'll be able to seek<a name="line.795"></a> -<span class="sourceLineNo">796</span> * past many Cells, especially if we know we need to go to the next block.</li><a name="line.796"></a> -<span class="sourceLineNo">797</span> * </ul><a name="line.797"></a> -<span class="sourceLineNo">798</span> * <p>A good proxy (best effort) to determine whether SKIP is better than SEEK is whether<a name="line.798"></a> -<span class="sourceLineNo">799</span> * we'll likely end up seeking to the next block (or past the next block) to get our next column.<a name="line.799"></a> -<span class="sourceLineNo">800</span> * Example:<a name="line.800"></a> -<span class="sourceLineNo">801</span> * <pre><a name="line.801"></a> -<span class="sourceLineNo">802</span> * | BLOCK 1 | BLOCK 2 |<a name="line.802"></a> -<span class="sourceLineNo">803</span> * | r1/c1, r1/c2, r1/c3 | r1/c4, r1/c5, r2/c1 |<a name="line.803"></a> -<span class="sourceLineNo">804</span> * ^ ^<a name="line.804"></a> -<span class="sourceLineNo">805</span> * | |<a name="line.805"></a> -<span class="sourceLineNo">806</span> * Next Index Key SEEK_NEXT_ROW (before r2/c1)<a name="line.806"></a> +<span class="sourceLineNo">051</span>import org.apache.hadoop.hbase.shaded.com.google.common.annotations.VisibleForTesting;<a name="line.51"></a> +<span class="sourceLineNo">052</span>import org.apache.hadoop.hbase.util.CollectionUtils;<a name="line.52"></a> +<span class="sourceLineNo">053</span>import org.apache.hadoop.hbase.util.EnvironmentEdgeManager;<a name="line.53"></a> +<span class="sourceLineNo">054</span><a name="line.54"></a> +<span class="sourceLineNo">055</span>/**<a name="line.55"></a> +<span class="sourceLineNo">056</span> * Scanner scans both the memstore and the Store. Coalesce KeyValue stream into List&lt;KeyValue&gt;<a name="line.56"></a> +<span class="sourceLineNo">057</span> * for a single row.<a name="line.57"></a> +<span class="sourceLineNo">058</span> * <p><a name="line.58"></a> +<span class="sourceLineNo">059</span> * The implementation is not thread safe. So there will be no race between next and close. The only<a name="line.59"></a> +<span class="sourceLineNo">060</span> * exception is updateReaders, it will be called in the memstore flush thread to indicate that there<a name="line.60"></a> +<span class="sourceLineNo">061</span> * is a flush.<a name="line.61"></a> +<span class="sourceLineNo">062</span> */<a name="line.62"></a> +<span class="sourceLineNo">063</span>@InterfaceAudience.Private<a name="line.63"></a> +<span class="sourceLineNo">064</span>public class StoreScanner extends NonReversedNonLazyKeyValueScanner<a name="line.64"></a> +<span class="sourceLineNo">065</span> implements KeyValueScanner, InternalScanner, ChangedReadersObserver {<a name="line.65"></a> +<span class="sourceLineNo">066</span> private static final Log LOG = LogFactory.getLog(StoreScanner.class);<a name="line.66"></a> +<span class="sourceLineNo">067</span> // In unit tests, the store could be null<a name="line.67"></a> +<span class="sourceLineNo">068</span> protected final Store store;<a name="line.68"></a> +<span class="sourceLineNo">069</span> private ScanQueryMatcher matcher;<a name="line.69"></a> +<span class="sourceLineNo">070</span> protected KeyValueHeap heap;<a name="line.70"></a> +<span class="sourceLineNo">071</span> private boolean cacheBlocks;<a name="line.71"></a> +<span class="sourceLineNo">072</span><a name="line.72"></a> +<span class="sourceLineNo">073</span> private long countPerRow = 0;<a name="line.73"></a> +<span class="sourceLineNo">074</span> private int storeLimit = -1;<a name="line.74"></a> +<span class="sourceLineNo">075</span> private int storeOffset = 0;<a name="line.75"></a> +<span class="sourceLineNo">076</span><a name="line.76"></a> +<span class="sourceLineNo">077</span> // Used to indicate that the scanner has closed (see HBASE-1107)<a name="line.77"></a> +<span class="sourceLineNo">078</span> // Do not need to be volatile because it's always accessed via synchronized methods<a name="line.78"></a> +<span class="sourceLineNo">079</span> private boolean closing = false;<a name="line.79"></a> +<span class="sourceLineNo">080</span> private final boolean get;<a name="line.80"></a> +<span class="sourceLineNo">081</span> private final boolean explicitColumnQuery;<a name="line.81"></a> +<span class="sourceLineNo">082</span> private final boolean useRowColBloom;<a name="line.82"></a> +<span class="sourceLineNo">083</span> /**<a name="line.83"></a> +<span class="sourceLineNo">084</span> * A flag that enables StoreFileScanner parallel-seeking<a name="line.84"></a> +<span class="sourceLineNo">085</span> */<a name="line.85"></a> +<span class="sourceLineNo">086</span> private boolean parallelSeekEnabled = false;<a name="line.86"></a> +<span class="sourceLineNo">087</span> private ExecutorService executor;<a name="line.87"></a> +<span class="sourceLineNo">088</span> private final Scan scan;<a name="line.88"></a> +<span class="sourceLineNo">089</span> private final long oldestUnexpiredTS;<a name="line.89"></a> +<span class="sourceLineNo">090</span> private final long now;<a name="line.90"></a> +<span class="sourceLineNo">091</span> private final int minVersions;<a name="line.91"></a> +<span class="sourceLineNo">092</span> private final long maxRowSize;<a name="line.92"></a> +<span class="sourceLineNo">093</span> private final long cellsPerHeartbeatCheck;<a name="line.93"></a> +<span class="sourceLineNo">094</span><a name="line.94"></a> +<span class="sourceLineNo">095</span> // 1) Collects all the KVHeap that are eagerly getting closed during the<a name="line.95"></a> +<span class="sourceLineNo">096</span> // course of a scan<a name="line.96"></a> +<span class="sourceLineNo">097</span> // 2) Collects the unused memstore scanners. If we close the memstore scanners<a name="line.97"></a> +<span class="sourceLineNo">098</span> // before sending data to client, the chunk may be reclaimed by other<a name="line.98"></a> +<span class="sourceLineNo">099</span> // updates and the data will be corrupt.<a name="line.99"></a> +<span class="sourceLineNo">100</span> private final List<KeyValueScanner> scannersForDelayedClose = new ArrayList<>();<a name="line.100"></a> +<span class="sourceLineNo">101</span><a name="line.101"></a> +<span class="sourceLineNo">102</span> /**<a name="line.102"></a> +<span class="sourceLineNo">103</span> * The number of KVs seen by the scanner. Includes explicitly skipped KVs, but not<a name="line.103"></a> +<span class="sourceLineNo">104</span> * KVs skipped via seeking to next row/column. TODO: estimate them?<a name="line.104"></a> +<span class="sourceLineNo">105</span> */<a name="line.105"></a> +<span class="sourceLineNo">106</span> private long kvsScanned = 0;<a name="line.106"></a> +<span class="sourceLineNo">107</span> protected Cell prevCell = null;<a name="line.107"></a> +<span class="sourceLineNo">108</span><a name="line.108"></a> +<span class="sourceLineNo">109</span> private final long preadMaxBytes;<a name="line.109"></a> +<span class="sourceLineNo">110</span> private long bytesRead;<a name="line.110"></a> +<span class="sourceLineNo">111</span><a name="line.111"></a> +<span class="sourceLineNo">112</span> /** We don't ever expect to change this, the constant is just for clarity. */<a name="line.112"></a> +<span class="sourceLineNo">113</span> static final boolean LAZY_SEEK_ENABLED_BY_DEFAULT = true;<a name="line.113"></a> +<span class="sourceLineNo">114</span> public static final String STORESCANNER_PARALLEL_SEEK_ENABLE =<a name="line.114"></a> +<span class="sourceLineNo">115</span> "hbase.storescanner.parallel.seek.enable";<a name="line.115"></a> +<span class="sourceLineNo">116</span><a name="line.116"></a> +<span class="sourceLineNo">117</span> /** Used during unit testing to ensure that lazy seek does save seek ops */<a name="line.117"></a> +<span class="sourceLineNo">118</span> private static boolean lazySeekEnabledGlobally = LAZY_SEEK_ENABLED_BY_DEFAULT;<a name="line.118"></a> +<span class="sourceLineNo">119</span><a name="line.119"></a> +<span class="sourceLineNo">120</span> /**<a name="line.120"></a> +<span class="sourceLineNo">121</span> * The number of cells scanned in between timeout checks. Specifying a larger value means that<a name="line.121"></a> +<span class="sourceLineNo">122</span> * timeout checks will occur less frequently. Specifying a small value will lead to more frequent<a name="line.122"></a> +<span class="sourceLineNo">123</span> * timeout checks.<a name="line.123"></a> +<span class="sourceLineNo">124</span> */<a name="line.124"></a> +<span class="sourceLineNo">125</span> public static final String HBASE_CELLS_SCANNED_PER_HEARTBEAT_CHECK =<a name="line.125"></a> +<span class="sourceLineNo">126</span> "hbase.cells.scanned.per.heartbeat.check";<a name="line.126"></a> +<span class="sourceLineNo">127</span><a name="line.127"></a> +<span class="sourceLineNo">128</span> /**<a name="line.128"></a> +<span class="sourceLineNo">129</span> * Default value of {@link #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> public static final long DEFAULT_HBASE_CELLS_SCANNED_PER_HEARTBEAT_CHECK = 10000;<a name="line.131"></a> +<span class="sourceLineNo">132</span><a name="line.132"></a> +<span class="sourceLineNo">133</span> /**<a name="line.133"></a> +<span class="sourceLineNo">134</span> * If the read type if Scan.ReadType.DEFAULT, we will start with pread, and if the kvs we scanned<a name="line.134"></a> +<span class="sourceLineNo">135</span> * reaches this limit, we will reopen the scanner with stream. The default value is 4 times of<a name="line.135"></a> +<span class="sourceLineNo">136</span> * block size for this store.<a name="line.136"></a> +<span class="sourceLineNo">137</span> */<a name="line.137"></a> +<span class="sourceLineNo">138</span> public static final String STORESCANNER_PREAD_MAX_BYTES = "hbase.storescanner.pread.max.bytes";<a name="line.138"></a> +<span class="sourceLineNo">139</span><a name="line.139"></a> +<span class="sourceLineNo">140</span> private final Scan.ReadType readType;<a name="line.140"></a> +<span class="sourceLineNo">141</span><a name="line.141"></a> +<span class="sourceLineNo">142</span> // A flag whether use pread for scan<a name="line.142"></a> +<span class="sourceLineNo">143</span> // it maybe changed if we use Scan.ReadType.DEFAULT and we have read lots of data.<a name="line.143"></a> +<span class="sourceLineNo">144</span> private boolean scanUsePread;<a name="line.144"></a> +<span class="sourceLineNo">145</span> // Indicates whether there was flush during the course of the scan<a name="line.145"></a> +<span class="sourceLineNo">146</span> private volatile boolean flushed = false;<a name="line.146"></a> +<span class="sourceLineNo">147</span> // generally we get one file from a flush<a name="line.147"></a> +<span class="sourceLineNo">148</span> private final List<StoreFile> flushedStoreFiles = new ArrayList<>(1);<a name="line.148"></a> +<span class="sourceLineNo">149</span> // Since CompactingMemstore is now default, we get three memstore scanners from a flush<a name="line.149"></a> +<span class="sourceLineNo">150</span> private final List<KeyValueScanner> memStoreScannersAfterFlush = new ArrayList<>(3);<a name="line.150"></a> +<span class="sourceLineNo">151</span> // The current list of scanners<a name="line.151"></a> +<span class="sourceLineNo">152</span> @VisibleForTesting<a name="line.152"></a> +<span class="sourceLineNo">153</span> final List<KeyValueScanner> currentScanners = new ArrayList<>();<a name="line.153"></a> +<span class="sourceLineNo">154</span> // flush update lock<a name="line.154"></a> +<span class="sourceLineNo">155</span> private final ReentrantLock flushLock = new ReentrantLock();<a name="line.155"></a> +<span class="sourceLineNo">156</span><a name="line.156"></a> +<span class="sourceLineNo">157</span> protected final long readPt;<a name="line.157"></a> +<span class="sourceLineNo">158</span> private boolean topChanged = false;<a name="line.158"></a> +<span class="sourceLineNo">159</span><a name="line.159"></a> +<span class="sourceLineNo">160</span> // used by the injection framework to test race between StoreScanner construction and compaction<a name="line.160"></a> +<span class="sourceLineNo">161</span> enum StoreScannerCompactionRace {<a name="line.161"></a> +<span class="sourceLineNo">162</span> BEFORE_SEEK,<a name="line.162"></a> +<span class="sourceLineNo">163</span> AFTER_SEEK,<a name="line.163"></a> +<span class="sourceLineNo">164</span> COMPACT_COMPLET
<TRUNCATED>
