You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@hbase.apache.org by gi...@apache.org on 2018/07/23 14:50:24 UTC
[10/31] hbase-site git commit: Published site at
4804483f7e55edf91a8e9d7ad30ad8239a96eaf3.
http://git-wip-us.apache.org/repos/asf/hbase-site/blob/b248f175/devapidocs/src-html/org/apache/hadoop/hbase/wal/WALSplitter.PipelineController.html
----------------------------------------------------------------------
diff --git a/devapidocs/src-html/org/apache/hadoop/hbase/wal/WALSplitter.PipelineController.html b/devapidocs/src-html/org/apache/hadoop/hbase/wal/WALSplitter.PipelineController.html
index 2d5c946..63e4b46 100644
--- a/devapidocs/src-html/org/apache/hadoop/hbase/wal/WALSplitter.PipelineController.html
+++ b/devapidocs/src-html/org/apache/hadoop/hbase/wal/WALSplitter.PipelineController.html
@@ -71,1836 +71,1860 @@
<span class="sourceLineNo">063</span>import org.apache.hadoop.hbase.client.Durability;<a name="line.63"></a>
<span class="sourceLineNo">064</span>import org.apache.hadoop.hbase.client.Mutation;<a name="line.64"></a>
<span class="sourceLineNo">065</span>import org.apache.hadoop.hbase.client.Put;<a name="line.65"></a>
-<span class="sourceLineNo">066</span>import org.apache.hadoop.hbase.coordination.SplitLogWorkerCoordination;<a name="line.66"></a>
-<span class="sourceLineNo">067</span>import org.apache.hadoop.hbase.io.HeapSize;<a name="line.67"></a>
-<span class="sourceLineNo">068</span>import org.apache.hadoop.hbase.log.HBaseMarkers;<a name="line.68"></a>
-<span class="sourceLineNo">069</span>import org.apache.hadoop.hbase.master.SplitLogManager;<a name="line.69"></a>
-<span class="sourceLineNo">070</span>import org.apache.hadoop.hbase.monitoring.MonitoredTask;<a name="line.70"></a>
-<span class="sourceLineNo">071</span>import org.apache.hadoop.hbase.monitoring.TaskMonitor;<a name="line.71"></a>
-<span class="sourceLineNo">072</span>import org.apache.hadoop.hbase.regionserver.HRegion;<a name="line.72"></a>
-<span class="sourceLineNo">073</span>import org.apache.hadoop.hbase.regionserver.LastSequenceId;<a name="line.73"></a>
-<span class="sourceLineNo">074</span>import org.apache.hadoop.hbase.regionserver.wal.AbstractFSWAL;<a name="line.74"></a>
-<span class="sourceLineNo">075</span>import org.apache.hadoop.hbase.regionserver.wal.WALCellCodec;<a name="line.75"></a>
-<span class="sourceLineNo">076</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.76"></a>
-<span class="sourceLineNo">077</span>import org.apache.hadoop.hbase.util.CancelableProgressable;<a name="line.77"></a>
-<span class="sourceLineNo">078</span>import org.apache.hadoop.hbase.util.ClassSize;<a name="line.78"></a>
-<span class="sourceLineNo">079</span>import org.apache.hadoop.hbase.util.FSUtils;<a name="line.79"></a>
-<span class="sourceLineNo">080</span>import org.apache.hadoop.hbase.util.Pair;<a name="line.80"></a>
-<span class="sourceLineNo">081</span>import org.apache.hadoop.hbase.util.Threads;<a name="line.81"></a>
-<span class="sourceLineNo">082</span>import org.apache.hadoop.hbase.wal.WAL.Entry;<a name="line.82"></a>
-<span class="sourceLineNo">083</span>import org.apache.hadoop.hbase.wal.WAL.Reader;<a name="line.83"></a>
-<span class="sourceLineNo">084</span>import org.apache.hadoop.hbase.wal.WALProvider.Writer;<a name="line.84"></a>
-<span class="sourceLineNo">085</span>import org.apache.hadoop.hbase.zookeeper.ZKSplitLog;<a name="line.85"></a>
-<span class="sourceLineNo">086</span>import org.apache.hadoop.io.MultipleIOException;<a name="line.86"></a>
-<span class="sourceLineNo">087</span>import org.apache.hadoop.ipc.RemoteException;<a name="line.87"></a>
-<span class="sourceLineNo">088</span>import org.apache.yetus.audience.InterfaceAudience;<a name="line.88"></a>
-<span class="sourceLineNo">089</span>import org.slf4j.Logger;<a name="line.89"></a>
-<span class="sourceLineNo">090</span>import org.slf4j.LoggerFactory;<a name="line.90"></a>
-<span class="sourceLineNo">091</span><a name="line.91"></a>
-<span class="sourceLineNo">092</span>import org.apache.hbase.thirdparty.com.google.common.annotations.VisibleForTesting;<a name="line.92"></a>
-<span class="sourceLineNo">093</span>import org.apache.hbase.thirdparty.com.google.common.base.Preconditions;<a name="line.93"></a>
-<span class="sourceLineNo">094</span>import org.apache.hbase.thirdparty.com.google.common.collect.Lists;<a name="line.94"></a>
-<span class="sourceLineNo">095</span>import org.apache.hbase.thirdparty.com.google.protobuf.TextFormat;<a name="line.95"></a>
-<span class="sourceLineNo">096</span>import org.apache.hbase.thirdparty.org.apache.commons.collections4.CollectionUtils;<a name="line.96"></a>
-<span class="sourceLineNo">097</span>import org.apache.hbase.thirdparty.org.apache.commons.collections4.MapUtils;<a name="line.97"></a>
-<span class="sourceLineNo">098</span><a name="line.98"></a>
-<span class="sourceLineNo">099</span>import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.WALEntry;<a name="line.99"></a>
-<span class="sourceLineNo">100</span>import org.apache.hadoop.hbase.shaded.protobuf.generated.ClientProtos.MutationProto.MutationType;<a name="line.100"></a>
-<span class="sourceLineNo">101</span>import org.apache.hadoop.hbase.shaded.protobuf.generated.ClusterStatusProtos.RegionStoreSequenceIds;<a name="line.101"></a>
-<span class="sourceLineNo">102</span>import org.apache.hadoop.hbase.shaded.protobuf.generated.ClusterStatusProtos.StoreSequenceId;<a name="line.102"></a>
-<span class="sourceLineNo">103</span>import org.apache.hadoop.hbase.shaded.protobuf.generated.HBaseProtos;<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> * This class is responsible for splitting up a bunch of regionserver commit log<a name="line.106"></a>
-<span class="sourceLineNo">107</span> * files that are no longer being written to, into new files, one per region, for<a name="line.107"></a>
-<span class="sourceLineNo">108</span> * recovering data on startup. Delete the old log files when finished.<a name="line.108"></a>
-<span class="sourceLineNo">109</span> */<a name="line.109"></a>
-<span class="sourceLineNo">110</span>@InterfaceAudience.Private<a name="line.110"></a>
-<span class="sourceLineNo">111</span>public class WALSplitter {<a name="line.111"></a>
-<span class="sourceLineNo">112</span> private static final Logger LOG = LoggerFactory.getLogger(WALSplitter.class);<a name="line.112"></a>
-<span class="sourceLineNo">113</span><a name="line.113"></a>
-<span class="sourceLineNo">114</span> /** By default we retry errors in splitting, rather than skipping. */<a name="line.114"></a>
-<span class="sourceLineNo">115</span> public static final boolean SPLIT_SKIP_ERRORS_DEFAULT = false;<a name="line.115"></a>
-<span class="sourceLineNo">116</span><a name="line.116"></a>
-<span class="sourceLineNo">117</span> // Parameters for split process<a name="line.117"></a>
-<span class="sourceLineNo">118</span> protected final Path walDir;<a name="line.118"></a>
-<span class="sourceLineNo">119</span> protected final FileSystem fs;<a name="line.119"></a>
-<span class="sourceLineNo">120</span> protected final Configuration conf;<a name="line.120"></a>
-<span class="sourceLineNo">121</span><a name="line.121"></a>
-<span class="sourceLineNo">122</span> // Major subcomponents of the split process.<a name="line.122"></a>
-<span class="sourceLineNo">123</span> // These are separated into inner classes to make testing easier.<a name="line.123"></a>
-<span class="sourceLineNo">124</span> OutputSink outputSink;<a name="line.124"></a>
-<span class="sourceLineNo">125</span> private EntryBuffers entryBuffers;<a name="line.125"></a>
-<span class="sourceLineNo">126</span><a name="line.126"></a>
-<span class="sourceLineNo">127</span> private SplitLogWorkerCoordination splitLogWorkerCoordination;<a name="line.127"></a>
-<span class="sourceLineNo">128</span> private final WALFactory walFactory;<a name="line.128"></a>
-<span class="sourceLineNo">129</span><a name="line.129"></a>
-<span class="sourceLineNo">130</span> private MonitoredTask status;<a name="line.130"></a>
-<span class="sourceLineNo">131</span><a name="line.131"></a>
-<span class="sourceLineNo">132</span> // For checking the latest flushed sequence id<a name="line.132"></a>
-<span class="sourceLineNo">133</span> protected final LastSequenceId sequenceIdChecker;<a name="line.133"></a>
-<span class="sourceLineNo">134</span><a name="line.134"></a>
-<span class="sourceLineNo">135</span> // Map encodedRegionName -> lastFlushedSequenceId<a name="line.135"></a>
-<span class="sourceLineNo">136</span> protected Map<String, Long> lastFlushedSequenceIds = new ConcurrentHashMap<>();<a name="line.136"></a>
-<span class="sourceLineNo">137</span><a name="line.137"></a>
-<span class="sourceLineNo">138</span> // Map encodedRegionName -> maxSeqIdInStores<a name="line.138"></a>
-<span class="sourceLineNo">139</span> protected Map<String, Map<byte[], Long>> regionMaxSeqIdInStores = new ConcurrentHashMap<>();<a name="line.139"></a>
-<span class="sourceLineNo">140</span><a name="line.140"></a>
-<span class="sourceLineNo">141</span> // the file being split currently<a name="line.141"></a>
-<span class="sourceLineNo">142</span> private FileStatus fileBeingSplit;<a name="line.142"></a>
-<span class="sourceLineNo">143</span><a name="line.143"></a>
-<span class="sourceLineNo">144</span> // if we limit the number of writers opened for sinking recovered edits<a name="line.144"></a>
-<span class="sourceLineNo">145</span> private final boolean splitWriterCreationBounded;<a name="line.145"></a>
-<span class="sourceLineNo">146</span><a name="line.146"></a>
-<span class="sourceLineNo">147</span> public final static String SPLIT_WRITER_CREATION_BOUNDED = "hbase.split.writer.creation.bounded";<a name="line.147"></a>
-<span class="sourceLineNo">148</span><a name="line.148"></a>
+<span class="sourceLineNo">066</span>import org.apache.hadoop.hbase.client.RegionInfo;<a name="line.66"></a>
+<span class="sourceLineNo">067</span>import org.apache.hadoop.hbase.coordination.SplitLogWorkerCoordination;<a name="line.67"></a>
+<span class="sourceLineNo">068</span>import org.apache.hadoop.hbase.io.HeapSize;<a name="line.68"></a>
+<span class="sourceLineNo">069</span>import org.apache.hadoop.hbase.log.HBaseMarkers;<a name="line.69"></a>
+<span class="sourceLineNo">070</span>import org.apache.hadoop.hbase.master.SplitLogManager;<a name="line.70"></a>
+<span class="sourceLineNo">071</span>import org.apache.hadoop.hbase.monitoring.MonitoredTask;<a name="line.71"></a>
+<span class="sourceLineNo">072</span>import org.apache.hadoop.hbase.monitoring.TaskMonitor;<a name="line.72"></a>
+<span class="sourceLineNo">073</span>import org.apache.hadoop.hbase.regionserver.HRegion;<a name="line.73"></a>
+<span class="sourceLineNo">074</span>import org.apache.hadoop.hbase.regionserver.LastSequenceId;<a name="line.74"></a>
+<span class="sourceLineNo">075</span>import org.apache.hadoop.hbase.regionserver.wal.AbstractFSWAL;<a name="line.75"></a>
+<span class="sourceLineNo">076</span>import org.apache.hadoop.hbase.regionserver.wal.WALCellCodec;<a name="line.76"></a>
+<span class="sourceLineNo">077</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.77"></a>
+<span class="sourceLineNo">078</span>import org.apache.hadoop.hbase.util.CancelableProgressable;<a name="line.78"></a>
+<span class="sourceLineNo">079</span>import org.apache.hadoop.hbase.util.ClassSize;<a name="line.79"></a>
+<span class="sourceLineNo">080</span>import org.apache.hadoop.hbase.util.FSUtils;<a name="line.80"></a>
+<span class="sourceLineNo">081</span>import org.apache.hadoop.hbase.util.Pair;<a name="line.81"></a>
+<span class="sourceLineNo">082</span>import org.apache.hadoop.hbase.util.Threads;<a name="line.82"></a>
+<span class="sourceLineNo">083</span>import org.apache.hadoop.hbase.wal.WAL.Entry;<a name="line.83"></a>
+<span class="sourceLineNo">084</span>import org.apache.hadoop.hbase.wal.WAL.Reader;<a name="line.84"></a>
+<span class="sourceLineNo">085</span>import org.apache.hadoop.hbase.wal.WALProvider.Writer;<a name="line.85"></a>
+<span class="sourceLineNo">086</span>import org.apache.hadoop.hbase.zookeeper.ZKSplitLog;<a name="line.86"></a>
+<span class="sourceLineNo">087</span>import org.apache.hadoop.io.MultipleIOException;<a name="line.87"></a>
+<span class="sourceLineNo">088</span>import org.apache.hadoop.ipc.RemoteException;<a name="line.88"></a>
+<span class="sourceLineNo">089</span>import org.apache.yetus.audience.InterfaceAudience;<a name="line.89"></a>
+<span class="sourceLineNo">090</span>import org.slf4j.Logger;<a name="line.90"></a>
+<span class="sourceLineNo">091</span>import org.slf4j.LoggerFactory;<a name="line.91"></a>
+<span class="sourceLineNo">092</span><a name="line.92"></a>
+<span class="sourceLineNo">093</span>import org.apache.hbase.thirdparty.com.google.common.annotations.VisibleForTesting;<a name="line.93"></a>
+<span class="sourceLineNo">094</span>import org.apache.hbase.thirdparty.com.google.common.base.Preconditions;<a name="line.94"></a>
+<span class="sourceLineNo">095</span>import org.apache.hbase.thirdparty.com.google.common.collect.Lists;<a name="line.95"></a>
+<span class="sourceLineNo">096</span>import org.apache.hbase.thirdparty.com.google.protobuf.TextFormat;<a name="line.96"></a>
+<span class="sourceLineNo">097</span>import org.apache.hbase.thirdparty.org.apache.commons.collections4.CollectionUtils;<a name="line.97"></a>
+<span class="sourceLineNo">098</span>import org.apache.hbase.thirdparty.org.apache.commons.collections4.MapUtils;<a name="line.98"></a>
+<span class="sourceLineNo">099</span><a name="line.99"></a>
+<span class="sourceLineNo">100</span>import org.apache.hadoop.hbase.shaded.protobuf.generated.AdminProtos.WALEntry;<a name="line.100"></a>
+<span class="sourceLineNo">101</span>import org.apache.hadoop.hbase.shaded.protobuf.generated.ClientProtos.MutationProto.MutationType;<a name="line.101"></a>
+<span class="sourceLineNo">102</span>import org.apache.hadoop.hbase.shaded.protobuf.generated.ClusterStatusProtos.RegionStoreSequenceIds;<a name="line.102"></a>
+<span class="sourceLineNo">103</span>import org.apache.hadoop.hbase.shaded.protobuf.generated.ClusterStatusProtos.StoreSequenceId;<a name="line.103"></a>
+<span class="sourceLineNo">104</span>import org.apache.hadoop.hbase.shaded.protobuf.generated.HBaseProtos;<a name="line.104"></a>
+<span class="sourceLineNo">105</span><a name="line.105"></a>
+<span class="sourceLineNo">106</span>/**<a name="line.106"></a>
+<span class="sourceLineNo">107</span> * This class is responsible for splitting up a bunch of regionserver commit log<a name="line.107"></a>
+<span class="sourceLineNo">108</span> * files that are no longer being written to, into new files, one per region, for<a name="line.108"></a>
+<span class="sourceLineNo">109</span> * recovering data on startup. Delete the old log files when finished.<a name="line.109"></a>
+<span class="sourceLineNo">110</span> */<a name="line.110"></a>
+<span class="sourceLineNo">111</span>@InterfaceAudience.Private<a name="line.111"></a>
+<span class="sourceLineNo">112</span>public class WALSplitter {<a name="line.112"></a>
+<span class="sourceLineNo">113</span> private static final Logger LOG = LoggerFactory.getLogger(WALSplitter.class);<a name="line.113"></a>
+<span class="sourceLineNo">114</span><a name="line.114"></a>
+<span class="sourceLineNo">115</span> /** By default we retry errors in splitting, rather than skipping. */<a name="line.115"></a>
+<span class="sourceLineNo">116</span> public static final boolean SPLIT_SKIP_ERRORS_DEFAULT = false;<a name="line.116"></a>
+<span class="sourceLineNo">117</span><a name="line.117"></a>
+<span class="sourceLineNo">118</span> // Parameters for split process<a name="line.118"></a>
+<span class="sourceLineNo">119</span> protected final Path walDir;<a name="line.119"></a>
+<span class="sourceLineNo">120</span> protected final FileSystem fs;<a name="line.120"></a>
+<span class="sourceLineNo">121</span> protected final Configuration conf;<a name="line.121"></a>
+<span class="sourceLineNo">122</span><a name="line.122"></a>
+<span class="sourceLineNo">123</span> // Major subcomponents of the split process.<a name="line.123"></a>
+<span class="sourceLineNo">124</span> // These are separated into inner classes to make testing easier.<a name="line.124"></a>
+<span class="sourceLineNo">125</span> OutputSink outputSink;<a name="line.125"></a>
+<span class="sourceLineNo">126</span> private EntryBuffers entryBuffers;<a name="line.126"></a>
+<span class="sourceLineNo">127</span><a name="line.127"></a>
+<span class="sourceLineNo">128</span> private SplitLogWorkerCoordination splitLogWorkerCoordination;<a name="line.128"></a>
+<span class="sourceLineNo">129</span> private final WALFactory walFactory;<a name="line.129"></a>
+<span class="sourceLineNo">130</span><a name="line.130"></a>
+<span class="sourceLineNo">131</span> private MonitoredTask status;<a name="line.131"></a>
+<span class="sourceLineNo">132</span><a name="line.132"></a>
+<span class="sourceLineNo">133</span> // For checking the latest flushed sequence id<a name="line.133"></a>
+<span class="sourceLineNo">134</span> protected final LastSequenceId sequenceIdChecker;<a name="line.134"></a>
+<span class="sourceLineNo">135</span><a name="line.135"></a>
+<span class="sourceLineNo">136</span> // Map encodedRegionName -> lastFlushedSequenceId<a name="line.136"></a>
+<span class="sourceLineNo">137</span> protected Map<String, Long> lastFlushedSequenceIds = new ConcurrentHashMap<>();<a name="line.137"></a>
+<span class="sourceLineNo">138</span><a name="line.138"></a>
+<span class="sourceLineNo">139</span> // Map encodedRegionName -> maxSeqIdInStores<a name="line.139"></a>
+<span class="sourceLineNo">140</span> protected Map<String, Map<byte[], Long>> regionMaxSeqIdInStores = new ConcurrentHashMap<>();<a name="line.140"></a>
+<span class="sourceLineNo">141</span><a name="line.141"></a>
+<span class="sourceLineNo">142</span> // the file being split currently<a name="line.142"></a>
+<span class="sourceLineNo">143</span> private FileStatus fileBeingSplit;<a name="line.143"></a>
+<span class="sourceLineNo">144</span><a name="line.144"></a>
+<span class="sourceLineNo">145</span> // if we limit the number of writers opened for sinking recovered edits<a name="line.145"></a>
+<span class="sourceLineNo">146</span> private final boolean splitWriterCreationBounded;<a name="line.146"></a>
+<span class="sourceLineNo">147</span><a name="line.147"></a>
+<span class="sourceLineNo">148</span> public final static String SPLIT_WRITER_CREATION_BOUNDED = "hbase.split.writer.creation.bounded";<a name="line.148"></a>
<span class="sourceLineNo">149</span><a name="line.149"></a>
-<span class="sourceLineNo">150</span> @VisibleForTesting<a name="line.150"></a>
-<span class="sourceLineNo">151</span> WALSplitter(final WALFactory factory, Configuration conf, Path walDir,<a name="line.151"></a>
-<span class="sourceLineNo">152</span> FileSystem fs, LastSequenceId idChecker,<a name="line.152"></a>
-<span class="sourceLineNo">153</span> SplitLogWorkerCoordination splitLogWorkerCoordination) {<a name="line.153"></a>
-<span class="sourceLineNo">154</span> this.conf = HBaseConfiguration.create(conf);<a name="line.154"></a>
-<span class="sourceLineNo">155</span> String codecClassName = conf<a name="line.155"></a>
-<span class="sourceLineNo">156</span> .get(WALCellCodec.WAL_CELL_CODEC_CLASS_KEY, WALCellCodec.class.getName());<a name="line.156"></a>
-<span class="sourceLineNo">157</span> this.conf.set(HConstants.RPC_CODEC_CONF_KEY, codecClassName);<a name="line.157"></a>
-<span class="sourceLineNo">158</span> this.walDir = walDir;<a name="line.158"></a>
-<span class="sourceLineNo">159</span> this.fs = fs;<a name="line.159"></a>
-<span class="sourceLineNo">160</span> this.sequenceIdChecker = idChecker;<a name="line.160"></a>
-<span class="sourceLineNo">161</span> this.splitLogWorkerCoordination = splitLogWorkerCoordination;<a name="line.161"></a>
-<span class="sourceLineNo">162</span><a name="line.162"></a>
-<span class="sourceLineNo">163</span> this.walFactory = factory;<a name="line.163"></a>
-<span class="sourceLineNo">164</span> PipelineController controller = new PipelineController();<a name="line.164"></a>
-<span class="sourceLineNo">165</span><a name="line.165"></a>
-<span class="sourceLineNo">166</span> this.splitWriterCreationBounded = conf.getBoolean(SPLIT_WRITER_CREATION_BOUNDED, false);<a name="line.166"></a>
-<span class="sourceLineNo">167</span><a name="line.167"></a>
-<span class="sourceLineNo">168</span> entryBuffers = new EntryBuffers(controller,<a name="line.168"></a>
-<span class="sourceLineNo">169</span> this.conf.getInt("hbase.regionserver.hlog.splitlog.buffersize", 128 * 1024 * 1024),<a name="line.169"></a>
-<span class="sourceLineNo">170</span> splitWriterCreationBounded);<a name="line.170"></a>
-<span class="sourceLineNo">171</span><a name="line.171"></a>
-<span class="sourceLineNo">172</span> int numWriterThreads = this.conf.getInt("hbase.regionserver.hlog.splitlog.writer.threads", 3);<a name="line.172"></a>
-<span class="sourceLineNo">173</span> if(splitWriterCreationBounded){<a name="line.173"></a>
-<span class="sourceLineNo">174</span> outputSink = new BoundedLogWriterCreationOutputSink(<a name="line.174"></a>
-<span class="sourceLineNo">175</span> controller, entryBuffers, numWriterThreads);<a name="line.175"></a>
-<span class="sourceLineNo">176</span> }else {<a name="line.176"></a>
-<span class="sourceLineNo">177</span> outputSink = new LogRecoveredEditsOutputSink(controller, entryBuffers, numWriterThreads);<a name="line.177"></a>
-<span class="sourceLineNo">178</span> }<a name="line.178"></a>
-<span class="sourceLineNo">179</span> }<a name="line.179"></a>
-<span class="sourceLineNo">180</span><a name="line.180"></a>
-<span class="sourceLineNo">181</span> /**<a name="line.181"></a>
-<span class="sourceLineNo">182</span> * Splits a WAL file into region's recovered-edits directory.<a name="line.182"></a>
-<span class="sourceLineNo">183</span> * This is the main entry point for distributed log splitting from SplitLogWorker.<a name="line.183"></a>
-<span class="sourceLineNo">184</span> * <p><a name="line.184"></a>
-<span class="sourceLineNo">185</span> * If the log file has N regions then N recovered.edits files will be produced.<a name="line.185"></a>
-<span class="sourceLineNo">186</span> * <p><a name="line.186"></a>
-<span class="sourceLineNo">187</span> * @return false if it is interrupted by the progress-able.<a name="line.187"></a>
-<span class="sourceLineNo">188</span> */<a name="line.188"></a>
-<span class="sourceLineNo">189</span> public static boolean splitLogFile(Path walDir, FileStatus logfile, FileSystem fs,<a name="line.189"></a>
-<span class="sourceLineNo">190</span> Configuration conf, CancelableProgressable reporter, LastSequenceId idChecker,<a name="line.190"></a>
-<span class="sourceLineNo">191</span> SplitLogWorkerCoordination splitLogWorkerCoordination, final WALFactory factory)<a name="line.191"></a>
-<span class="sourceLineNo">192</span> throws IOException {<a name="line.192"></a>
-<span class="sourceLineNo">193</span> WALSplitter s = new WALSplitter(factory, conf, walDir, fs, idChecker,<a name="line.193"></a>
-<span class="sourceLineNo">194</span> splitLogWorkerCoordination);<a name="line.194"></a>
-<span class="sourceLineNo">195</span> return s.splitLogFile(logfile, reporter);<a name="line.195"></a>
-<span class="sourceLineNo">196</span> }<a name="line.196"></a>
-<span class="sourceLineNo">197</span><a name="line.197"></a>
-<span class="sourceLineNo">198</span> // A wrapper to split one log folder using the method used by distributed<a name="line.198"></a>
-<span class="sourceLineNo">199</span> // log splitting. Used by tools and unit tests. It should be package private.<a name="line.199"></a>
-<span class="sourceLineNo">200</span> // It is public only because TestWALObserver is in a different package,<a name="line.200"></a>
-<span class="sourceLineNo">201</span> // which uses this method to do log splitting.<a name="line.201"></a>
-<span class="sourceLineNo">202</span> @VisibleForTesting<a name="line.202"></a>
-<span class="sourceLineNo">203</span> public static List<Path> split(Path rootDir, Path logDir, Path oldLogDir,<a name="line.203"></a>
-<span class="sourceLineNo">204</span> FileSystem fs, Configuration conf, final WALFactory factory) throws IOException {<a name="line.204"></a>
-<span class="sourceLineNo">205</span> final FileStatus[] logfiles = SplitLogManager.getFileList(conf,<a name="line.205"></a>
-<span class="sourceLineNo">206</span> Collections.singletonList(logDir), null);<a name="line.206"></a>
-<span class="sourceLineNo">207</span> List<Path> splits = new ArrayList<>();<a name="line.207"></a>
-<span class="sourceLineNo">208</span> if (ArrayUtils.isNotEmpty(logfiles)) {<a name="line.208"></a>
-<span class="sourceLineNo">209</span> for (FileStatus logfile: logfiles) {<a name="line.209"></a>
-<span class="sourceLineNo">210</span> WALSplitter s = new WALSplitter(factory, conf, rootDir, fs, null, null);<a name="line.210"></a>
-<span class="sourceLineNo">211</span> if (s.splitLogFile(logfile, null)) {<a name="line.211"></a>
-<span class="sourceLineNo">212</span> finishSplitLogFile(rootDir, oldLogDir, logfile.getPath(), conf);<a name="line.212"></a>
-<span class="sourceLineNo">213</span> if (s.outputSink.splits != null) {<a name="line.213"></a>
-<span class="sourceLineNo">214</span> splits.addAll(s.outputSink.splits);<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> if (!fs.delete(logDir, true)) {<a name="line.219"></a>
-<span class="sourceLineNo">220</span> throw new IOException("Unable to delete src dir: " + logDir);<a name="line.220"></a>
-<span class="sourceLineNo">221</span> }<a name="line.221"></a>
-<span class="sourceLineNo">222</span> return splits;<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> /**<a name="line.225"></a>
-<span class="sourceLineNo">226</span> * log splitting implementation, splits one log file.<a name="line.226"></a>
-<span class="sourceLineNo">227</span> * @param logfile should be an actual log file.<a name="line.227"></a>
-<span class="sourceLineNo">228</span> */<a name="line.228"></a>
-<span class="sourceLineNo">229</span> @VisibleForTesting<a name="line.229"></a>
-<span class="sourceLineNo">230</span> boolean splitLogFile(FileStatus logfile, CancelableProgressable reporter) throws IOException {<a name="line.230"></a>
-<span class="sourceLineNo">231</span> Preconditions.checkState(status == null);<a name="line.231"></a>
-<span class="sourceLineNo">232</span> Preconditions.checkArgument(logfile.isFile(),<a name="line.232"></a>
-<span class="sourceLineNo">233</span> "passed in file status is for something other than a regular file.");<a name="line.233"></a>
-<span class="sourceLineNo">234</span> boolean isCorrupted = false;<a name="line.234"></a>
-<span class="sourceLineNo">235</span> boolean skipErrors = conf.getBoolean("hbase.hlog.split.skip.errors",<a name="line.235"></a>
-<span class="sourceLineNo">236</span> SPLIT_SKIP_ERRORS_DEFAULT);<a name="line.236"></a>
-<span class="sourceLineNo">237</span> int interval = conf.getInt("hbase.splitlog.report.interval.loglines", 1024);<a name="line.237"></a>
-<span class="sourceLineNo">238</span> Path logPath = logfile.getPath();<a name="line.238"></a>
-<span class="sourceLineNo">239</span> boolean outputSinkStarted = false;<a name="line.239"></a>
-<span class="sourceLineNo">240</span> boolean progress_failed = false;<a name="line.240"></a>
-<span class="sourceLineNo">241</span> int editsCount = 0;<a name="line.241"></a>
-<span class="sourceLineNo">242</span> int editsSkipped = 0;<a name="line.242"></a>
-<span class="sourceLineNo">243</span><a name="line.243"></a>
-<span class="sourceLineNo">244</span> status = TaskMonitor.get().createStatus(<a name="line.244"></a>
-<span class="sourceLineNo">245</span> "Splitting log file " + logfile.getPath() + "into a temporary staging area.");<a name="line.245"></a>
-<span class="sourceLineNo">246</span> Reader logFileReader = null;<a name="line.246"></a>
-<span class="sourceLineNo">247</span> this.fileBeingSplit = logfile;<a name="line.247"></a>
-<span class="sourceLineNo">248</span> try {<a name="line.248"></a>
-<span class="sourceLineNo">249</span> long logLength = logfile.getLen();<a name="line.249"></a>
-<span class="sourceLineNo">250</span> LOG.info("Splitting WAL={}, length={}", logPath, logLength);<a name="line.250"></a>
-<span class="sourceLineNo">251</span> status.setStatus("Opening log file");<a name="line.251"></a>
-<span class="sourceLineNo">252</span> if (reporter != null && !reporter.progress()) {<a name="line.252"></a>
-<span class="sourceLineNo">253</span> progress_failed = true;<a name="line.253"></a>
-<span class="sourceLineNo">254</span> return false;<a name="line.254"></a>
-<span class="sourceLineNo">255</span> }<a name="line.255"></a>
-<span class="sourceLineNo">256</span> logFileReader = getReader(logfile, skipErrors, reporter);<a name="line.256"></a>
-<span class="sourceLineNo">257</span> if (logFileReader == null) {<a name="line.257"></a>
-<span class="sourceLineNo">258</span> LOG.warn("Nothing to split in WAL={}", logPath);<a name="line.258"></a>
-<span class="sourceLineNo">259</span> return true;<a name="line.259"></a>
-<span class="sourceLineNo">260</span> }<a name="line.260"></a>
-<span class="sourceLineNo">261</span> int numOpenedFilesBeforeReporting = conf.getInt("hbase.splitlog.report.openedfiles", 3);<a name="line.261"></a>
-<span class="sourceLineNo">262</span> int numOpenedFilesLastCheck = 0;<a name="line.262"></a>
-<span class="sourceLineNo">263</span> outputSink.setReporter(reporter);<a name="line.263"></a>
-<span class="sourceLineNo">264</span> outputSink.startWriterThreads();<a name="line.264"></a>
-<span class="sourceLineNo">265</span> outputSinkStarted = true;<a name="line.265"></a>
-<span class="sourceLineNo">266</span> Entry entry;<a name="line.266"></a>
-<span class="sourceLineNo">267</span> Long lastFlushedSequenceId = -1L;<a name="line.267"></a>
-<span class="sourceLineNo">268</span> while ((entry = getNextLogLine(logFileReader, logPath, skipErrors)) != null) {<a name="line.268"></a>
-<span class="sourceLineNo">269</span> byte[] region = entry.getKey().getEncodedRegionName();<a name="line.269"></a>
-<span class="sourceLineNo">270</span> String encodedRegionNameAsStr = Bytes.toString(region);<a name="line.270"></a>
-<span class="sourceLineNo">271</span> lastFlushedSequenceId = lastFlushedSequenceIds.get(encodedRegionNameAsStr);<a name="line.271"></a>
-<span class="sourceLineNo">272</span> if (lastFlushedSequenceId == null) {<a name="line.272"></a>
-<span class="sourceLineNo">273</span> if (sequenceIdChecker != null) {<a name="line.273"></a>
-<span class="sourceLineNo">274</span> RegionStoreSequenceIds ids = sequenceIdChecker.getLastSequenceId(region);<a name="line.274"></a>
-<span class="sourceLineNo">275</span> Map<byte[], Long> maxSeqIdInStores = new TreeMap<>(Bytes.BYTES_COMPARATOR);<a name="line.275"></a>
-<span class="sourceLineNo">276</span> for (StoreSequenceId storeSeqId : ids.getStoreSequenceIdList()) {<a name="line.276"></a>
-<span class="sourceLineNo">277</span> maxSeqIdInStores.put(storeSeqId.getFamilyName().toByteArray(),<a name="line.277"></a>
-<span class="sourceLineNo">278</span> storeSeqId.getSequenceId());<a name="line.278"></a>
-<span class="sourceLineNo">279</span> }<a name="line.279"></a>
-<span class="sourceLineNo">280</span> regionMaxSeqIdInStores.put(encodedRegionNameAsStr, maxSeqIdInStores);<a name="line.280"></a>
-<span class="sourceLineNo">281</span> lastFlushedSequenceId = ids.getLastFlushedSequenceId();<a name="line.281"></a>
-<span class="sourceLineNo">282</span> if (LOG.isDebugEnabled()) {<a name="line.282"></a>
-<span class="sourceLineNo">283</span> LOG.debug("DLS Last flushed sequenceid for " + encodedRegionNameAsStr + ": " +<a name="line.283"></a>
-<span class="sourceLineNo">284</span> TextFormat.shortDebugString(ids));<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> if (lastFlushedSequenceId == null) {<a name="line.287"></a>
-<span class="sourceLineNo">288</span> lastFlushedSequenceId = -1L;<a name="line.288"></a>
-<span class="sourceLineNo">289</span> }<a name="line.289"></a>
-<span class="sourceLineNo">290</span> lastFlushedSequenceIds.put(encodedRegionNameAsStr, lastFlushedSequenceId);<a name="line.290"></a>
-<span class="sourceLineNo">291</span> }<a name="line.291"></a>
-<span class="sourceLineNo">292</span> if (lastFlushedSequenceId >= entry.getKey().getSequenceId()) {<a name="line.292"></a>
-<span class="sourceLineNo">293</span> editsSkipped++;<a name="line.293"></a>
-<span class="sourceLineNo">294</span> continue;<a name="line.294"></a>
-<span class="sourceLineNo">295</span> }<a name="line.295"></a>
-<span class="sourceLineNo">296</span> // Don't send Compaction/Close/Open region events to recovered edit type sinks.<a name="line.296"></a>
-<span class="sourceLineNo">297</span> if (entry.getEdit().isMetaEdit() && !outputSink.keepRegionEvent(entry)) {<a name="line.297"></a>
-<span class="sourceLineNo">298</span> editsSkipped++;<a name="line.298"></a>
-<span class="sourceLineNo">299</span> continue;<a name="line.299"></a>
-<span class="sourceLineNo">300</span> }<a name="line.300"></a>
-<span class="sourceLineNo">301</span> entryBuffers.appendEntry(entry);<a name="line.301"></a>
-<span class="sourceLineNo">302</span> editsCount++;<a name="line.302"></a>
-<span class="sourceLineNo">303</span> int moreWritersFromLastCheck = this.getNumOpenWriters() - numOpenedFilesLastCheck;<a name="line.303"></a>
-<span class="sourceLineNo">304</span> // If sufficient edits have passed, check if we should report progress.<a name="line.304"></a>
-<span class="sourceLineNo">305</span> if (editsCount % interval == 0<a name="line.305"></a>
-<span class="sourceLineNo">306</span> || moreWritersFromLastCheck > numOpenedFilesBeforeReporting) {<a name="line.306"></a>
-<span class="sourceLineNo">307</span> numOpenedFilesLastCheck = this.getNumOpenWriters();<a name="line.307"></a>
-<span class="sourceLineNo">308</span> String countsStr = (editsCount - (editsSkipped + outputSink.getSkippedEdits()))<a name="line.308"></a>
-<span class="sourceLineNo">309</span> + " edits, skipped " + editsSkipped + " edits.";<a name="line.309"></a>
-<span class="sourceLineNo">310</span> status.setStatus("Split " + countsStr);<a name="line.310"></a>
-<span class="sourceLineNo">311</span> if (reporter != null && !reporter.progress()) {<a name="line.311"></a>
-<span class="sourceLineNo">312</span> progress_failed = true;<a name="line.312"></a>
-<span class="sourceLineNo">313</span> return false;<a name="line.313"></a>
-<span class="sourceLineNo">314</span> }<a name="line.314"></a>
-<span class="sourceLineNo">315</span> }<a name="line.315"></a>
-<span class="sourceLineNo">316</span> }<a name="line.316"></a>
-<span class="sourceLineNo">317</span> } catch (InterruptedException ie) {<a name="line.317"></a>
-<span class="sourceLineNo">318</span> IOException iie = new InterruptedIOException();<a name="line.318"></a>
-<span class="sourceLineNo">319</span> iie.initCause(ie);<a name="line.319"></a>
-<span class="sourceLineNo">320</span> throw iie;<a name="line.320"></a>
-<span class="sourceLineNo">321</span> } catch (CorruptedLogFileException e) {<a name="line.321"></a>
-<span class="sourceLineNo">322</span> LOG.warn("Could not parse, corrupted WAL={}", logPath, e);<a name="line.322"></a>
-<span class="sourceLineNo">323</span> if (splitLogWorkerCoordination != null) {<a name="line.323"></a>
-<span class="sourceLineNo">324</span> // Some tests pass in a csm of null.<a name="line.324"></a>
-<span class="sourceLineNo">325</span> splitLogWorkerCoordination.markCorrupted(walDir, logfile.getPath().getName(), fs);<a name="line.325"></a>
-<span class="sourceLineNo">326</span> } else {<a name="line.326"></a>
-<span class="sourceLineNo">327</span> // for tests only<a name="line.327"></a>
-<span class="sourceLineNo">328</span> ZKSplitLog.markCorrupted(walDir, logfile.getPath().getName(), fs);<a name="line.328"></a>
-<span class="sourceLineNo">329</span> }<a name="line.329"></a>
-<span class="sourceLineNo">330</span> isCorrupted = true;<a name="line.330"></a>
-<span class="sourceLineNo">331</span> } catch (IOException e) {<a name="line.331"></a>
-<span class="sourceLineNo">332</span> e = e instanceof RemoteException ? ((RemoteException) e).unwrapRemoteException() : e;<a name="line.332"></a>
-<span class="sourceLineNo">333</span> throw e;<a name="line.333"></a>
-<span class="sourceLineNo">334</span> } finally {<a name="line.334"></a>
-<span class="sourceLineNo">335</span> LOG.debug("Finishing writing output logs and closing down");<a name="line.335"></a>
-<span class="sourceLineNo">336</span> try {<a name="line.336"></a>
-<span class="sourceLineNo">337</span> if (null != logFileReader) {<a name="line.337"></a>
-<span class="sourceLineNo">338</span> logFileReader.close();<a name="line.338"></a>
-<span class="sourceLineNo">339</span> }<a name="line.339"></a>
-<span class="sourceLineNo">340</span> } catch (IOException exception) {<a name="line.340"></a>
-<span class="sourceLineNo">341</span> LOG.warn("Could not close WAL reader", exception);<a name="line.341"></a>
-<span class="sourceLineNo">342</span> }<a name="line.342"></a>
-<span class="sourceLineNo">343</span> try {<a name="line.343"></a>
-<span class="sourceLineNo">344</span> if (outputSinkStarted) {<a name="line.344"></a>
-<span class="sourceLineNo">345</span> // Set progress_failed to true as the immediate following statement will reset its value<a name="line.345"></a>
-<span class="sourceLineNo">346</span> // when finishWritingAndClose() throws exception, progress_failed has the right value<a name="line.346"></a>
-<span class="sourceLineNo">347</span> progress_failed = true;<a name="line.347"></a>
-<span class="sourceLineNo">348</span> progress_failed = outputSink.finishWritingAndClose() == null;<a name="line.348"></a>
-<span class="sourceLineNo">349</span> }<a name="line.349"></a>
-<span class="sourceLineNo">350</span> } finally {<a name="line.350"></a>
-<span class="sourceLineNo">351</span> String msg =<a name="line.351"></a>
-<span class="sourceLineNo">352</span> "Processed " + editsCount + " edits across " + outputSink.getNumberOfRecoveredRegions()<a name="line.352"></a>
-<span class="sourceLineNo">353</span> + " regions; edits skipped=" + editsSkipped + "; log file=" + logPath +<a name="line.353"></a>
-<span class="sourceLineNo">354</span> ", length=" + logfile.getLen() + // See if length got updated post lease recovery<a name="line.354"></a>
-<span class="sourceLineNo">355</span> ", corrupted=" + isCorrupted + ", progress failed=" + progress_failed;<a name="line.355"></a>
-<span class="sourceLineNo">356</span> LOG.info(msg);<a name="line.356"></a>
-<span class="sourceLineNo">357</span> status.markComplete(msg);<a name="line.357"></a>
-<span class="sourceLineNo">358</span> }<a name="line.358"></a>
-<span class="sourceLineNo">359</span> }<a name="line.359"></a>
-<span class="sourceLineNo">360</span> return !progress_failed;<a name="line.360"></a>
-<span class="sourceLineNo">361</span> }<a name="line.361"></a>
-<span class="sourceLineNo">362</span><a name="line.362"></a>
-<span class="sourceLineNo">363</span> /**<a name="line.363"></a>
-<span class="sourceLineNo">364</span> * Completes the work done by splitLogFile by archiving logs<a name="line.364"></a>
-<span class="sourceLineNo">365</span> * <p><a name="line.365"></a>
-<span class="sourceLineNo">366</span> * It is invoked by SplitLogManager once it knows that one of the<a name="line.366"></a>
-<span class="sourceLineNo">367</span> * SplitLogWorkers have completed the splitLogFile() part. If the master<a name="line.367"></a>
-<span class="sourceLineNo">368</span> * crashes then this function might get called multiple times.<a name="line.368"></a>
-<span class="sourceLineNo">369</span> * <p><a name="line.369"></a>
-<span class="sourceLineNo">370</span> * @param logfile<a name="line.370"></a>
-<span class="sourceLineNo">371</span> * @param conf<a name="line.371"></a>
-<span class="sourceLineNo">372</span> * @throws IOException<a name="line.372"></a>
-<span class="sourceLineNo">373</span> */<a name="line.373"></a>
-<span class="sourceLineNo">374</span> public static void finishSplitLogFile(String logfile,<a name="line.374"></a>
-<span class="sourceLineNo">375</span> Configuration conf) throws IOException {<a name="line.375"></a>
-<span class="sourceLineNo">376</span> Path rootdir = FSUtils.getWALRootDir(conf);<a name="line.376"></a>
-<span class="sourceLineNo">377</span> Path oldLogDir = new Path(rootdir, HConstants.HREGION_OLDLOGDIR_NAME);<a name="line.377"></a>
-<span class="sourceLineNo">378</span> Path logPath;<a name="line.378"></a>
-<span class="sourceLineNo">379</span> if (FSUtils.isStartingWithPath(rootdir, logfile)) {<a name="line.379"></a>
-<span class="sourceLineNo">380</span> logPath = new Path(logfile);<a name="line.380"></a>
-<span class="sourceLineNo">381</span> } else {<a name="line.381"></a>
-<span class="sourceLineNo">382</span> logPath = new Path(rootdir, logfile);<a name="line.382"></a>
-<span class="sourceLineNo">383</span> }<a name="line.383"></a>
-<span class="sourceLineNo">384</span> finishSplitLogFile(rootdir, oldLogDir, logPath, conf);<a name="line.384"></a>
-<span class="sourceLineNo">385</span> }<a name="line.385"></a>
-<span class="sourceLineNo">386</span><a name="line.386"></a>
-<span class="sourceLineNo">387</span> private static void finishSplitLogFile(Path rootdir, Path oldLogDir,<a name="line.387"></a>
-<span class="sourceLineNo">388</span> Path logPath, Configuration conf) throws IOException {<a name="line.388"></a>
-<span class="sourceLineNo">389</span> List<Path> processedLogs = new ArrayList<>();<a name="line.389"></a>
-<span class="sourceLineNo">390</span> List<Path> corruptedLogs = new ArrayList<>();<a name="line.390"></a>
-<span class="sourceLineNo">391</span> FileSystem fs;<a name="line.391"></a>
-<span class="sourceLineNo">392</span> fs = rootdir.getFileSystem(conf);<a name="line.392"></a>
-<span class="sourceLineNo">393</span> if (ZKSplitLog.isCorrupted(rootdir, logPath.getName(), fs)) {<a name="line.393"></a>
-<span class="sourceLineNo">394</span> corruptedLogs.add(logPath);<a name="line.394"></a>
-<span class="sourceLineNo">395</span> } else {<a name="line.395"></a>
-<span class="sourceLineNo">396</span> processedLogs.add(logPath);<a name="line.396"></a>
-<span class="sourceLineNo">397</span> }<a name="line.397"></a>
-<span class="sourceLineNo">398</span> archiveLogs(corruptedLogs, processedLogs, oldLogDir, fs, conf);<a name="line.398"></a>
-<span class="sourceLineNo">399</span> Path stagingDir = ZKSplitLog.getSplitLogDir(rootdir, logPath.getName());<a name="line.399"></a>
-<span class="sourceLineNo">400</span> fs.delete(stagingDir, true);<a name="line.400"></a>
-<span class="sourceLineNo">401</span> }<a name="line.401"></a>
-<span class="sourceLineNo">402</span><a name="line.402"></a>
-<span class="sourceLineNo">403</span> /**<a name="line.403"></a>
-<span class="sourceLineNo">404</span> * Moves processed logs to a oldLogDir after successful processing Moves<a name="line.404"></a>
-<span class="sourceLineNo">405</span> * corrupted logs (any log that couldn't be successfully parsed to corruptDir<a name="line.405"></a>
-<span class="sourceLineNo">406</span> * (.corrupt) for later investigation<a name="line.406"></a>
-<span class="sourceLineNo">407</span> *<a name="line.407"></a>
-<span class="sourceLineNo">408</span> * @param corruptedLogs<a name="line.408"></a>
-<span class="sourceLineNo">409</span> * @param processedLogs<a name="line.409"></a>
-<span class="sourceLineNo">410</span> * @param oldLogDir<a name="line.410"></a>
-<span class="sourceLineNo">411</span> * @param fs<a name="line.411"></a>
-<span class="sourceLineNo">412</span> * @param conf<a name="line.412"></a>
-<span class="sourceLineNo">413</span> * @throws IOException<a name="line.413"></a>
-<span class="sourceLineNo">414</span> */<a name="line.414"></a>
-<span class="sourceLineNo">415</span> private static void archiveLogs(<a name="line.415"></a>
-<span class="sourceLineNo">416</span> final List<Path> corruptedLogs,<a name="line.416"></a>
-<span class="sourceLineNo">417</span> final List<Path> processedLogs, final Path oldLogDir,<a name="line.417"></a>
-<span class="sourceLineNo">418</span> final FileSystem fs, final Configuration conf) throws IOException {<a name="line.418"></a>
-<span class="sourceLineNo">419</span> final Path corruptDir = new Path(FSUtils.getWALRootDir(conf), HConstants.CORRUPT_DIR_NAME);<a name="line.419"></a>
-<span class="sourceLineNo">420</span> if (conf.get("hbase.regionserver.hlog.splitlog.corrupt.dir") != null) {<a name="line.420"></a>
-<span class="sourceLineNo">421</span> LOG.warn("hbase.regionserver.hlog.splitlog.corrupt.dir is deprecated. Default to {}",<a name="line.421"></a>
-<span class="sourceLineNo">422</span> corruptDir);<a name="line.422"></a>
-<span class="sourceLineNo">423</span> }<a name="line.423"></a>
-<span class="sourceLineNo">424</span> if (!fs.mkdirs(corruptDir)) {<a name="line.424"></a>
-<span class="sourceLineNo">425</span> LOG.info("Unable to mkdir {}", corruptDir);<a name="line.425"></a>
-<span class="sourceLineNo">426</span> }<a name="line.426"></a>
-<span class="sourceLineNo">427</span> fs.mkdirs(oldLogDir);<a name="line.427"></a>
-<span class="sourceLineNo">428</span><a name="line.428"></a>
-<span class="sourceLineNo">429</span> // this method can get restarted or called multiple times for archiving<a name="line.429"></a>
-<span class="sourceLineNo">430</span> // the same log files.<a name="line.430"></a>
-<span class="sourceLineNo">431</span> for (Path corrupted : corruptedLogs) {<a name="line.431"></a>
-<span class="sourceLineNo">432</span> Path p = new Path(corruptDir, corrupted.getName());<a name="line.432"></a>
-<span class="sourceLineNo">433</span> if (fs.exists(corrupted)) {<a name="line.433"></a>
-<span class="sourceLineNo">434</span> if (!fs.rename(corrupted, p)) {<a name="line.434"></a>
-<span class="sourceLineNo">435</span> LOG.warn("Unable to move corrupted log {} to {}", corrupted, p);<a name="line.435"></a>
-<span class="sourceLineNo">436</span> } else {<a name="line.436"></a>
-<span class="sourceLineNo">437</span> LOG.warn("Moved corrupted log {} to {}", corrupted, p);<a name="line.437"></a>
-<span class="sourceLineNo">438</span> }<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> for (Path p : processedLogs) {<a name="line.442"></a>
-<span class="sourceLineNo">443</span> Path newPath = AbstractFSWAL.getWALArchivePath(oldLogDir, p);<a name="line.443"></a>
-<span class="sourceLineNo">444</span> if (fs.exists(p)) {<a name="line.444"></a>
-<span class="sourceLineNo">445</span> if (!FSUtils.renameAndSetModifyTime(fs, p, newPath)) {<a name="line.445"></a>
-<span class="sourceLineNo">446</span> LOG.warn("Unable to move {} to {}", p, newPath);<a name="line.446"></a>
-<span class="sourceLineNo">447</span> } else {<a name="line.447"></a>
-<span class="sourceLineNo">448</span> LOG.info("Archived processed log {} to {}", p, newPath);<a name="line.448"></a>
-<span class="sourceLineNo">449</span> }<a name="line.449"></a>
-<span class="sourceLineNo">450</span> }<a name="line.450"></a>
-<span class="sourceLineNo">451</span> }<a name="line.451"></a>
-<span class="sourceLineNo">452</span> }<a name="line.452"></a>
-<span class="sourceLineNo">453</span><a name="line.453"></a>
-<span class="sourceLineNo">454</span> /**<a name="line.454"></a>
-<span class="sourceLineNo">455</span> * Path to a file under RECOVERED_EDITS_DIR directory of the region found in<a name="line.455"></a>
-<span class="sourceLineNo">456</span> * <code>logEntry</code> named for the sequenceid in the passed<a name="line.456"></a>
-<span class="sourceLineNo">457</span> * <code>logEntry</code>: e.g. /hbase/some_table/2323432434/recovered.edits/2332.<a name="line.457"></a>
-<span class="sourceLineNo">458</span> * This method also ensures existence of RECOVERED_EDITS_DIR under the region<a name="line.458"></a>
-<span class="sourceLineNo">459</span> * creating it if necessary.<a name="line.459"></a>
-<span class="sourceLineNo">460</span> * @param logEntry<a name="line.460"></a>
-<span class="sourceLineNo">461</span> * @param fileNameBeingSplit the file being split currently. Used to generate tmp file name.<a name="line.461"></a>
-<span class="sourceLineNo">462</span> * @param conf<a name="line.462"></a>
-<span class="sourceLineNo">463</span> * @return Path to file into which to dump split log edits.<a name="line.463"></a>
-<span class="sourceLineNo">464</span> * @throws IOException<a name="line.464"></a>
-<span class="sourceLineNo">465</span> */<a name="line.465"></a>
-<span class="sourceLineNo">466</span> @SuppressWarnings("deprecation")<a name="line.466"></a>
-<span class="sourceLineNo">467</span> @VisibleForTesting<a name="line.467"></a>
-<span class="sourceLineNo">468</span> static Path getRegionSplitEditsPath(final Entry logEntry, String fileNameBeingSplit,<a name="line.468"></a>
-<span class="sourceLineNo">469</span> Configuration conf)<a name="line.469"></a>
-<span class="sourceLineNo">470</span> throws IOException {<a name="line.470"></a>
-<span class="sourceLineNo">471</span> FileSystem fs = FileSystem.get(conf);<a name="line.471"></a>
-<span class="sourceLineNo">472</span> Path rootDir = FSUtils.getRootDir(conf);<a name="line.472"></a>
-<span class="sourceLineNo">473</span> Path tableDir = FSUtils.getTableDir(rootDir, logEntry.getKey().getTableName());<a name="line.473"></a>
-<span class="sourceLineNo">474</span> String encodedRegionName = Bytes.toString(logEntry.getKey().getEncodedRegionName());<a name="line.474"></a>
-<span class="sourceLineNo">475</span> Path regiondir = HRegion.getRegionDir(tableDir, encodedRegionName);<a name="line.475"></a>
-<span class="sourceLineNo">476</span> Path dir = getRegionDirRecoveredEditsDir(regiondir);<a name="line.476"></a>
-<span class="sourceLineNo">477</span><a name="line.477"></a>
-<span class="sourceLineNo">478</span> if (!fs.exists(regiondir)) {<a name="line.478"></a>
-<span class="sourceLineNo">479</span> LOG.info("This region's directory does not exist: {}."<a name="line.479"></a>
-<span class="sourceLineNo">480</span> + "It is very likely that it was already split so it is "<a name="line.480"></a>
-<span class="sourceLineNo">481</span> + "safe to discard those edits.", regiondir);<a name="line.481"></a>
-<span class="sourceLineNo">482</span> return null;<a name="line.482"></a>
-<span class="sourceLineNo">483</span> }<a name="line.483"></a>
-<span class="sourceLineNo">484</span> if (fs.exists(dir) && fs.isFile(dir)) {<a name="line.484"></a>
-<span class="sourceLineNo">485</span> Path tmp = new Path("/tmp");<a name="line.485"></a>
-<span class="sourceLineNo">486</span> if (!fs.exists(tmp)) {<a name="line.486"></a>
-<span class="sourceLineNo">487</span> fs.mkdirs(tmp);<a name="line.487"></a>
-<span class="sourceLineNo">488</span> }<a name="line.488"></a>
-<span class="sourceLineNo">489</span> tmp = new Path(tmp,<a name="line.489"></a>
-<span class="sourceLineNo">490</span> HConstants.RECOVERED_EDITS_DIR + "_" + encodedRegionName);<a name="line.490"></a>
-<span class="sourceLineNo">491</span> LOG.warn("Found existing old file: {}. It could be some "<a name="line.491"></a>
-<span class="sourceLineNo">492</span> + "leftover of an old installation. It should be a folder instead. "<a name="line.492"></a>
-<span class="sourceLineNo">493</span> + "So moving it to {}", dir, tmp);<a name="line.493"></a>
-<span class="sourceLineNo">494</span> if (!fs.rename(dir, tmp)) {<a name="line.494"></a>
-<span class="sourceLineNo">495</span> LOG.warn("Failed to sideline old file {}", dir);<a name="line.495"></a>
-<span class="sourceLineNo">496</span> }<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> if (!fs.exists(dir) && !fs.mkdirs(dir)) {<a name="line.499"></a>
-<span class="sourceLineNo">500</span> LOG.warn("mkdir failed on {}", dir);<a name="line.500"></a>
-<span class="sourceLineNo">501</span> }<a name="line.501"></a>
-<span class="sourceLineNo">502</span> // Append fileBeingSplit to prevent name conflict since we may have duplicate wal entries now.<a name="line.502"></a>
-<span class="sourceLineNo">503</span> // Append file name ends with RECOVERED_LOG_TMPFILE_SUFFIX to ensure<a name="line.503"></a>
-<span class="sourceLineNo">504</span> // region's replayRecoveredEdits will not delete it<a name="line.504"></a>
-<span class="sourceLineNo">505</span> String fileName = formatRecoveredEditsFileName(logEntry.getKey().getSequenceId());<a name="line.505"></a>
-<span class="sourceLineNo">506</span> fileName = getTmpRecoveredEditsFileName(fileName + "-" + fileNameBeingSplit);<a name="line.506"></a>
-<span class="sourceLineNo">507</span> return new Path(dir, fileName);<a name="line.507"></a>
-<span class="sourceLineNo">508</span> }<a name="line.508"></a>
-<span class="sourceLineNo">509</span><a name="line.509"></a>
-<span class="sourceLineNo">510</span> private static String getTmpRecoveredEditsFileName(String fileName) {<a name="line.510"></a>
-<span class="sourceLineNo">511</span> return fileName + RECOVERED_LOG_TMPFILE_SUFFIX;<a name="line.511"></a>
-<span class="sourceLineNo">512</span> }<a name="line.512"></a>
-<span class="sourceLineNo">513</span><a name="line.513"></a>
-<span class="sourceLineNo">514</span> /**<a name="line.514"></a>
-<span class="sourceLineNo">515</span> * Get the completed recovered edits file path, renaming it to be by last edit<a name="line.515"></a>
-<span class="sourceLineNo">516</span> * in the file from its first edit. Then we could use the name to skip<a name="line.516"></a>
-<span class="sourceLineNo">517</span> * recovered edits when doing {@link HRegion#replayRecoveredEditsIfAny}.<a name="line.517"></a>
-<span class="sourceLineNo">518</span> * @param srcPath<a name="line.518"></a>
-<span class="sourceLineNo">519</span> * @param maximumEditLogSeqNum<a name="line.519"></a>
-<span class="sourceLineNo">520</span> * @return dstPath take file's last edit log seq num as the name<a name="line.520"></a>
-<span class="sourceLineNo">521</span> */<a name="line.521"></a>
-<span class="sourceLineNo">522</span> private static Path getCompletedRecoveredEditsFilePath(Path srcPath,<a name="line.522"></a>
-<span class="sourceLineNo">523</span> long maximumEditLogSeqNum) {<a name="line.523"></a>
-<span class="sourceLineNo">524</span> String fileName = formatRecoveredEditsFileName(maximumEditLogSeqNum);<a name="line.524"></a>
-<span class="sourceLineNo">525</span> return new Path(srcPath.getParent(), fileName);<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> @VisibleForTesting<a name="line.528"></a>
-<span class="sourceLineNo">529</span> static String formatRecoveredEditsFileName(final long seqid) {<a name="line.529"></a>
-<span class="sourceLineNo">530</span> return String.format("%019d", seqid);<a name="line.530"></a>
-<span class="sourceLineNo">531</span> }<a name="line.531"></a>
-<span class="sourceLineNo">532</span><a name="line.532"></a>
-<span class="sourceLineNo">533</span> private static final Pattern EDITFILES_NAME_PATTERN = Pattern.compile("-?[0-9]+");<a name="line.533"></a>
-<span class="sourceLineNo">534</span> private static final String RECOVERED_LOG_TMPFILE_SUFFIX = ".temp";<a name="line.534"></a>
-<span class="sourceLineNo">535</span><a name="line.535"></a>
-<span class="sourceLineNo">536</span> /**<a name="line.536"></a>
-<span class="sourceLineNo">537</span> * @param regiondir<a name="line.537"></a>
-<span class="sourceLineNo">538</span> * This regions directory in the filesystem.<a name="line.538"></a>
-<span class="sourceLineNo">539</span> * @return The directory that holds recovered edits files for the region<a name="line.539"></a>
-<span class="sourceLineNo">540</span> * <code>regiondir</code><a name="line.540"></a>
-<span class="sourceLineNo">541</span> */<a name="line.541"></a>
-<span class="sourceLineNo">542</span> public static Path getRegionDirRecoveredEditsDir(final Path regiondir) {<a name="line.542"></a>
-<span class="sourceLineNo">543</span> return new Path(regiondir, HConstants.RECOVERED_EDITS_DIR);<a name="line.543"></a>
-<span class="sourceLineNo">544</span> }<a name="line.544"></a>
-<span class="sourceLineNo">545</span><a name="line.545"></a>
-<span class="sourceLineNo">546</span> /**<a name="line.546"></a>
-<span class="sourceLineNo">547</span> * Returns sorted set of edit files made by splitter, excluding files<a name="line.547"></a>
-<span class="sourceLineNo">548</span> * with '.temp' suffix.<a name="line.548"></a>
-<span class="sourceLineNo">549</span> *<a name="line.549"></a>
-<span class="sourceLineNo">550</span> * @param fs<a name="line.550"></a>
-<span class="sourceLineNo">551</span> * @param regiondir<a name="line.551"></a>
-<span class="sourceLineNo">552</span> * @return Files in passed <code>regiondir</code> as a sorted set.<a name="line.552"></a>
-<span class="sourceLineNo">553</span> * @throws IOException<a name="line.553"></a>
+<span class="sourceLineNo">150</span><a name="line.150"></a>
+<span class="sourceLineNo">151</span> @VisibleForTesting<a name="line.151"></a>
+<span class="sourceLineNo">152</span> WALSplitter(final WALFactory factory, Configuration conf, Path walDir,<a name="line.152"></a>
+<span class="sourceLineNo">153</span> FileSystem fs, LastSequenceId idChecker,<a name="line.153"></a>
+<span class="sourceLineNo">154</span> SplitLogWorkerCoordination splitLogWorkerCoordination) {<a name="line.154"></a>
+<span class="sourceLineNo">155</span> this.conf = HBaseConfiguration.create(conf);<a name="line.155"></a>
+<span class="sourceLineNo">156</span> String codecClassName = conf<a name="line.156"></a>
+<span class="sourceLineNo">157</span> .get(WALCellCodec.WAL_CELL_CODEC_CLASS_KEY, WALCellCodec.class.getName());<a name="line.157"></a>
+<span class="sourceLineNo">158</span> this.conf.set(HConstants.RPC_CODEC_CONF_KEY, codecClassName);<a name="line.158"></a>
+<span class="sourceLineNo">159</span> this.walDir = walDir;<a name="line.159"></a>
+<span class="sourceLineNo">160</span> this.fs = fs;<a name="line.160"></a>
+<span class="sourceLineNo">161</span> this.sequenceIdChecker = idChecker;<a name="line.161"></a>
+<span class="sourceLineNo">162</span> this.splitLogWorkerCoordination = splitLogWorkerCoordination;<a name="line.162"></a>
+<span class="sourceLineNo">163</span><a name="line.163"></a>
+<span class="sourceLineNo">164</span> this.walFactory = factory;<a name="line.164"></a>
+<span class="sourceLineNo">165</span> PipelineController controller = new PipelineController();<a name="line.165"></a>
+<span class="sourceLineNo">166</span><a name="line.166"></a>
+<span class="sourceLineNo">167</span> this.splitWriterCreationBounded = conf.getBoolean(SPLIT_WRITER_CREATION_BOUNDED, false);<a name="line.167"></a>
+<span class="sourceLineNo">168</span><a name="line.168"></a>
+<span class="sourceLineNo">169</span> entryBuffers = new EntryBuffers(controller,<a name="line.169"></a>
+<span class="sourceLineNo">170</span> this.conf.getInt("hbase.regionserver.hlog.splitlog.buffersize", 128 * 1024 * 1024),<a name="line.170"></a>
+<span class="sourceLineNo">171</span> splitWriterCreationBounded);<a name="line.171"></a>
+<span class="sourceLineNo">172</span><a name="line.172"></a>
+<span class="sourceLineNo">173</span> int numWriterThreads = this.conf.getInt("hbase.regionserver.hlog.splitlog.writer.threads", 3);<a name="line.173"></a>
+<span class="sourceLineNo">174</span> if(splitWriterCreationBounded){<a name="line.174"></a>
+<span class="sourceLineNo">175</span> outputSink = new BoundedLogWriterCreationOutputSink(<a name="line.175"></a>
+<span class="sourceLineNo">176</span> controller, entryBuffers, numWriterThreads);<a name="line.176"></a>
+<span class="sourceLineNo">177</span> }else {<a name="line.177"></a>
+<span class="sourceLineNo">178</span> outputSink = new LogRecoveredEditsOutputSink(controller, entryBuffers, numWriterThreads);<a name="line.178"></a>
+<span class="sourceLineNo">179</span> }<a name="line.179"></a>
+<span class="sourceLineNo">180</span> }<a name="line.180"></a>
+<span class="sourceLineNo">181</span><a name="line.181"></a>
+<span class="sourceLineNo">182</span> /**<a name="line.182"></a>
+<span class="sourceLineNo">183</span> * Splits a WAL file into region's recovered-edits directory.<a name="line.183"></a>
+<span class="sourceLineNo">184</span> * This is the main entry point for distributed log splitting from SplitLogWorker.<a name="line.184"></a>
+<span class="sourceLineNo">185</span> * <p><a name="line.185"></a>
+<span class="sourceLineNo">186</span> * If the log file has N regions then N recovered.edits files will be produced.<a name="line.186"></a>
+<span class="sourceLineNo">187</span> * <p><a name="line.187"></a>
+<span class="sourceLineNo">188</span> * @return false if it is interrupted by the progress-able.<a name="line.188"></a>
+<span class="sourceLineNo">189</span> */<a name="line.189"></a>
+<span class="sourceLineNo">190</span> public static boolean splitLogFile(Path walDir, FileStatus logfile, FileSystem fs,<a name="line.190"></a>
+<span class="sourceLineNo">191</span> Configuration conf, CancelableProgressable reporter, LastSequenceId idChecker,<a name="line.191"></a>
+<span class="sourceLineNo">192</span> SplitLogWorkerCoordination splitLogWorkerCoordination, final WALFactory factory)<a name="line.192"></a>
+<span class="sourceLineNo">193</span> throws IOException {<a name="line.193"></a>
+<span class="sourceLineNo">194</span> WALSplitter s = new WALSplitter(factory, conf, walDir, fs, idChecker,<a name="line.194"></a>
+<span class="sourceLineNo">195</span> splitLogWorkerCoordination);<a name="line.195"></a>
+<span class="sourceLineNo">196</span> return s.splitLogFile(logfile, reporter);<a name="line.196"></a>
+<span class="sourceLineNo">197</span> }<a name="line.197"></a>
+<span class="sourceLineNo">198</span><a name="line.198"></a>
+<span class="sourceLineNo">199</span> // A wrapper to split one log folder using the method used by distributed<a name="line.199"></a>
+<span class="sourceLineNo">200</span> // log splitting. Used by tools and unit tests. It should be package private.<a name="line.200"></a>
+<span class="sourceLineNo">201</span> // It is public only because TestWALObserver is in a different package,<a name="line.201"></a>
+<span class="sourceLineNo">202</span> // which uses this method to do log splitting.<a name="line.202"></a>
+<span class="sourceLineNo">203</span> @VisibleForTesting<a name="line.203"></a>
+<span class="sourceLineNo">204</span> public static List<Path> split(Path rootDir, Path logDir, Path oldLogDir,<a name="line.204"></a>
+<span class="sourceLineNo">205</span> FileSystem fs, Configuration conf, final WALFactory factory) throws IOException {<a name="line.205"></a>
+<span class="sourceLineNo">206</span> final FileStatus[] logfiles = SplitLogManager.getFileList(conf,<a name="line.206"></a>
+<span class="sourceLineNo">207</span> Collections.singletonList(logDir), null);<a name="line.207"></a>
+<span class="sourceLineNo">208</span> List<Path> splits = new ArrayList<>();<a name="line.208"></a>
+<span class="sourceLineNo">209</span> if (ArrayUtils.isNotEmpty(logfiles)) {<a name="line.209"></a>
+<span class="sourceLineNo">210</span> for (FileStatus logfile: logfiles) {<a name="line.210"></a>
+<span class="sourceLineNo">211</span> WALSplitter s = new WALSplitter(factory, conf, rootDir, fs, null, null);<a name="line.211"></a>
+<span class="sourceLineNo">212</span> if (s.splitLogFile(logfile, null)) {<a name="line.212"></a>
+<span class="sourceLineNo">213</span> finishSplitLogFile(rootDir, oldLogDir, logfile.getPath(), conf);<a name="line.213"></a>
+<span class="sourceLineNo">214</span> if (s.outputSink.splits != null) {<a name="line.214"></a>
+<span class="sourceLineNo">215</span> splits.addAll(s.outputSink.splits);<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> if (!fs.delete(logDir, true)) {<a name="line.220"></a>
+<span class="sourceLineNo">221</span> throw new IOException("Unable to delete src dir: " + logDir);<a name="line.221"></a>
+<span class="sourceLineNo">222</span> }<a name="line.222"></a>
+<span class="sourceLineNo">223</span> return splits;<a name="line.223"></a>
+<span class="sourceLineNo">224</span> }<a name="line.224"></a>
+<span class="sourceLineNo">225</span><a name="line.225"></a>
+<span class="sourceLineNo">226</span> /**<a name="line.226"></a>
+<span class="sourceLineNo">227</span> * log splitting implementation, splits one log file.<a name="line.227"></a>
+<span class="sourceLineNo">228</span> * @param logfile should be an actual log file.<a name="line.228"></a>
+<span class="sourceLineNo">229</span> */<a name="line.229"></a>
+<span class="sourceLineNo">230</span> @VisibleForTesting<a name="line.230"></a>
+<span class="sourceLineNo">231</span> boolean splitLogFile(FileStatus logfile, CancelableProgressable reporter) throws IOException {<a name="line.231"></a>
+<span class="sourceLineNo">232</span> Preconditions.checkState(status == null);<a name="line.232"></a>
+<span class="sourceLineNo">233</span> Preconditions.checkArgument(logfile.isFile(),<a name="line.233"></a>
+<span class="sourceLineNo">234</span> "passed in file status is for something other than a regular file.");<a name="line.234"></a>
+<span class="sourceLineNo">235</span> boolean isCorrupted = false;<a name="line.235"></a>
+<span class="sourceLineNo">236</span> boolean skipErrors = conf.getBoolean("hbase.hlog.split.skip.errors",<a name="line.236"></a>
+<span class="sourceLineNo">237</span> SPLIT_SKIP_ERRORS_DEFAULT);<a name="line.237"></a>
+<span class="sourceLineNo">238</span> int interval = conf.getInt("hbase.splitlog.report.interval.loglines", 1024);<a name="line.238"></a>
+<span class="sourceLineNo">239</span> Path logPath = logfile.getPath();<a name="line.239"></a>
+<span class="sourceLineNo">240</span> boolean outputSinkStarted = false;<a name="line.240"></a>
+<span class="sourceLineNo">241</span> boolean progress_failed = false;<a name="line.241"></a>
+<span class="sourceLineNo">242</span> int editsCount = 0;<a name="line.242"></a>
+<span class="sourceLineNo">243</span> int editsSkipped = 0;<a name="line.243"></a>
+<span class="sourceLineNo">244</span><a name="line.244"></a>
+<span class="sourceLineNo">245</span> status = TaskMonitor.get().createStatus(<a name="line.245"></a>
+<span class="sourceLineNo">246</span> "Splitting log file " + logfile.getPath() + "into a temporary staging area.");<a name="line.246"></a>
+<span class="sourceLineNo">247</span> Reader logFileReader = null;<a name="line.247"></a>
+<span class="sourceLineNo">248</span> this.fileBeingSplit = logfile;<a name="line.248"></a>
+<span class="sourceLineNo">249</span> try {<a name="line.249"></a>
+<span class="sourceLineNo">250</span> long logLength = logfile.getLen();<a name="line.250"></a>
+<span class="sourceLineNo">251</span> LOG.info("Splitting WAL={}, length={}", logPath, logLength);<a name="line.251"></a>
+<span class="sourceLineNo">252</span> status.setStatus("Opening log file");<a name="line.252"></a>
+<span class="sourceLineNo">253</span> if (reporter != null && !reporter.progress()) {<a name="line.253"></a>
+<span class="sourceLineNo">254</span> progress_failed = true;<a name="line.254"></a>
+<span class="sourceLineNo">255</span> return false;<a name="line.255"></a>
+<span class="sourceLineNo">256</span> }<a name="line.256"></a>
+<span class="sourceLineNo">257</span> logFileReader = getReader(logfile, skipErrors, reporter);<a name="line.257"></a>
+<span class="sourceLineNo">258</span> if (logFileReader == null) {<a name="line.258"></a>
+<span class="sourceLineNo">259</span> LOG.warn("Nothing to split in WAL={}", logPath);<a name="line.259"></a>
+<span class="sourceLineNo">260</span> return true;<a name="line.260"></a>
+<span class="sourceLineNo">261</span> }<a name="line.261"></a>
+<span class="sourceLineNo">262</span> int numOpenedFilesBeforeReporting = conf.getInt("hbase.splitlog.report.openedfiles", 3);<a name="line.262"></a>
+<span class="sourceLineNo">263</span> int numOpenedFilesLastCheck = 0;<a name="line.263"></a>
+<span class="sourceLineNo">264</span> outputSink.setReporter(reporter);<a name="line.264"></a>
+<span class="sourceLineNo">265</span> outputSink.startWriterThreads();<a name="line.265"></a>
+<span class="sourceLineNo">266</span> outputSinkStarted = true;<a name="line.266"></a>
+<span class="sourceLineNo">267</span> Entry entry;<a name="line.267"></a>
+<span class="sourceLineNo">268</span> Long lastFlushedSequenceId = -1L;<a name="line.268"></a>
+<span class="sourceLineNo">269</span> while ((entry = getNextLogLine(logFileReader, logPath, skipErrors)) != null) {<a name="line.269"></a>
+<span class="sourceLineNo">270</span> byte[] region = entry.getKey().getEncodedRegionName();<a name="line.270"></a>
+<span class="sourceLineNo">271</span> String encodedRegionNameAsStr = Bytes.toString(region);<a name="line.271"></a>
+<span class="sourceLineNo">272</span> lastFlushedSequenceId = lastFlushedSequenceIds.get(encodedRegionNameAsStr);<a name="line.272"></a>
+<span class="sourceLineNo">273</span> if (lastFlushedSequenceId == null) {<a name="line.273"></a>
+<span class="sourceLineNo">274</span> if (sequenceIdChecker != null) {<a name="line.274"></a>
+<span class="sourceLineNo">275</span> RegionStoreSequenceIds ids = sequenceIdChecker.getLastSequenceId(region);<a name="line.275"></a>
+<span class="sourceLineNo">276</span> Map<byte[], Long> maxSeqIdInStores = new TreeMap<>(Bytes.BYTES_COMPARATOR);<a name="line.276"></a>
+<span class="sourceLineNo">277</span> for (StoreSequenceId storeSeqId : ids.getStoreSequenceIdList()) {<a name="line.277"></a>
+<span class="sourceLineNo">278</span> maxSeqIdInStores.put(storeSeqId.getFamilyName().toByteArray(),<a name="line.278"></a>
+<span class="sourceLineNo">279</span> storeSeqId.getSequenceId());<a name="line.279"></a>
+<span class="sourceLineNo">280</span> }<a name="line.280"></a>
+<span class="sourceLineNo">281</span> regionMaxSeqIdInStores.put(encodedRegionNameAsStr, maxSeqIdInStores);<a name="line.281"></a>
+<span class="sourceLineNo">282</span> lastFlushedSequenceId = ids.getLastFlushedSequenceId();<a name="line.282"></a>
+<span class="sourceLineNo">283</span> if (LOG.isDebugEnabled()) {<a name="line.283"></a>
+<span class="sourceLineNo">284</span> LOG.debug("DLS Last flushed sequenceid for " + encodedRegionNameAsStr + ": " +<a name="line.284"></a>
+<span class="sourceLineNo">285</span> TextFormat.shortDebugString(ids));<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> if (lastFlushedSequenceId == null) {<a name="line.288"></a>
+<span class="sourceLineNo">289</span> lastFlushedSequenceId = -1L;<a name="line.289"></a>
+<span class="sourceLineNo">290</span> }<a name="line.290"></a>
+<span class="sourceLineNo">291</span> lastFlushedSequenceIds.put(encodedRegionNameAsStr, lastFlushedSequenceId);<a name="line.291"></a>
+<span class="sourceLineNo">292</span> }<a name="line.292"></a>
+<span class="sourceLineNo">293</span> if (lastFlushedSequenceId >= entry.getKey().getSequenceId()) {<a name="line.293"></a>
+<span class="sourceLineNo">294</span> editsSkipped++;<a name="line.294"></a>
+<span class="sourceLineNo">295</span> continue;<a name="line.295"></a>
+<span class="sourceLineNo">296</span> }<a name="line.296"></a>
+<span class="sourceLineNo">297</span> // Don't send Compaction/Close/Open region events to recovered edit type sinks.<a name="line.297"></a>
+<span class="sourceLineNo">298</span> if (entry.getEdit().isMetaEdit() && !outputSink.keepRegionEvent(entry)) {<a name="line.298"></a>
+<span class="sourceLineNo">299</span> editsSkipped++;<a name="line.299"></a>
+<span class="sourceLineNo">300</span> continue;<a name="line.300"></a>
+<span class="sourceLineNo">301</span> }<a name="line.301"></a>
+<span class="sourceLineNo">302</span> entryBuffers.appendEntry(entry);<a name="line.302"></a>
+<span class="sourceLineNo">303</span> editsCount++;<a name="line.303"></a>
+<span class="sourceLineNo">304</span> int moreWritersFromLastCheck = this.getNumOpenWriters() - numOpenedFilesLastCheck;<a name="line.304"></a>
+<span class="sourceLineNo">305</span> // If sufficient edits have passed, check if we should report progress.<a name="line.305"></a>
+<span class="sourceLineNo">306</span> if (editsCount % interval == 0<a name="line.306"></a>
+<span class="sourceLineNo">307</span> || moreWritersFromLastCheck > numOpenedFilesBeforeReporting) {<a name="line.307"></a>
+<span class="sourceLineNo">308</span> numOpenedFilesLastCheck = this.getNumOpenWriters();<a name="line.308"></a>
+<span class="sourceLineNo">309</span> String countsStr = (editsCount - (editsSkipped + outputSink.getSkippedEdits()))<a name="line.309"></a>
+<span class="sourceLineNo">310</span> + " edits, skipped " + editsSkipped + " edits.";<a name="line.310"></a>
+<span class="sourceLineNo">311</span> status.setStatus("Split " + countsStr);<a name="line.311"></a>
+<span class="sourceLineNo">312</span> if (reporter != null && !reporter.progress()) {<a name="line.312"></a>
+<span class="sourceLineNo">313</span> progress_failed = true;<a name="line.313"></a>
+<span class="sourceLineNo">314</span> return false;<a name="line.314"></a>
+<span class="sourceLineNo">315</span> }<a name="line.315"></a>
+<span class="sourceLineNo">316</span> }<a name="line.316"></a>
+<span class="sourceLineNo">317</span> }<a name="line.317"></a>
+<span class="sourceLineNo">318</span> } catch (InterruptedException ie) {<a name="line.318"></a>
+<span class="sourceLineNo">319</span> IOException iie = new InterruptedIOException();<a name="line.319"></a>
+<span class="sourceLineNo">320</span> iie.initCause(ie);<a name="line.320"></a>
+<span class="sourceLineNo">321</span> throw iie;<a name="line.321"></a>
+<span class="sourceLineNo">322</span> } catch (CorruptedLogFileException e) {<a name="line.322"></a>
+<span class="sourceLineNo">323</span> LOG.warn("Could not parse, corrupted WAL={}", logPath, e);<a name="line.323"></a>
+<span class="sourceLineNo">324</span> if (splitLogWorkerCoordination != null) {<a name="line.324"></a>
+<span class="sourceLineNo">325</span> // Some tests pass in a csm of null.<a name="line.325"></a>
+<span class="sourceLineNo">326</span> splitLogWorkerCoordination.markCorrupted(walDir, logfile.getPath().getName(), fs);<a name="line.326"></a>
+<span class="sourceLineNo">327</span> } else {<a name="line.327"></a>
+<span class="sourceLineNo">328</span> // for tests only<a name="line.328"></a>
+<span class="sourceLineNo">329</span> ZKSplitLog.markCorrupted(walDir, logfile.getPath().getName(), fs);<a name="line.329"></a>
+<span class="sourceLineNo">330</span> }<a name="line.330"></a>
+<span class="sourceLineNo">331</span> isCorrupted = true;<a name="line.331"></a>
+<span class="sourceLineNo">332</span> } catch (IOException e) {<a name="line.332"></a>
+<span class="sourceLineNo">333</span> e = e instanceof RemoteException ? ((RemoteException) e).unwrapRemoteException() : e;<a name="line.333"></a>
+<span class="sourceLineNo">334</span> throw e;<a name="line.334"></a>
+<span class="sourceLineNo">335</span> } finally {<a name="line.335"></a>
+<span class="sourceLineNo">336</span> LOG.debug("Finishing writing output logs and closing down");<a name="line.336"></a>
+<span class="sourceLineNo">337</span> try {<a name="line.337"></a>
+<span class="sourceLineNo">338</span> if (null != logFileReader) {<a name="line.338"></a>
+<span class="sourceLineNo">339</span> logFileReader.close();<a name="line.339"></a>
+<span class="sourceLineNo">340</span> }<a name="line.340"></a>
+<span class="sourceLineNo">341</span> } catch (IOException exception) {<a name="line.341"></a>
+<span class="sourceLineNo">342</span> LOG.warn("Could not close WAL reader", exception);<a name="line.342"></a>
+<span class="sourceLineNo">343</span> }<a name="line.343"></a>
+<span class="sourceLineNo">344</span> try {<a name="line.344"></a>
+<span class="sourceLineNo">345</span> if (outputSinkStarted) {<a name="line.345"></a>
+<span class="sourceLineNo">346</span> // Set progress_failed to true as the immediate following statement will reset its value<a name="line.346"></a>
+<span class="sourceLineNo">347</span> // when finishWritingAndClose() throws exception, progress_failed has the right value<a name="line.347"></a>
+<span class="sourceLineNo">348</span> progress_failed = true;<a name="line.348"></a>
+<span class="sourceLineNo">349</span> progress_failed = outputSink.finishWritingAndClose() == null;<a name="line.349"></a>
+<span class="sourceLineNo">350</span> }<a name="line.350"></a>
+<span class="sourceLineNo">351</span> } finally {<a name="line.351"></a>
+<span class="sourceLineNo">352</span> String msg =<a name="line.352"></a>
+<span class="sourceLineNo">353</span> "Processed " + editsCount + " edits across " + outputSink.getNumberOfRecoveredRegions()<a name="line.353"></a>
+<span class="sourceLineNo">354</span> + " regions; edits skipped=" + editsSkipped + "; log file=" + logPath +<a name="line.354"></a>
+<span class="sourceLineNo">355</span> ", length=" + logfile.getLen() + // See if length got updated post lease recovery<a name="line.355"></a>
+<span class="sourceLineNo">356</span> ", corrupted=" + isCorrupted + ", progress failed=" + progress_failed;<a name="line.356"></a>
+<span class="sourceLineNo">357</span> LOG.info(msg);<a name="line.357"></a>
+<span class="sourceLineNo">358</span> status.markComplete(msg);<a name="line.358"></a>
+<span class="sourceLineNo">359</span> }<a name="line.359"></a>
+<span class="sourceLineNo">360</span> }<a name="line.360"></a>
+<span class="sourceLineNo">361</span> return !progress_failed;<a name="line.361"></a>
+<span class="sourceLineNo">362</span> }<a name="line.362"></a>
+<span class="sourceLineNo">363</span><a name="line.363"></a>
+<span class="sourceLineNo">364</span> /**<a name="line.364"></a>
+<span class="sourceLineNo">365</span> * Completes the work done by splitLogFile by archiving logs<a name="line.365"></a>
+<span class="sourceLineNo">366</span> * <p><a name="line.366"></a>
+<span class="sourceLineNo">367</span> * It is invoked by SplitLogManager once it knows that one of the<a name="line.367"></a>
+<span class="sourceLineNo">368</span> * SplitLogWorkers have completed the splitLogFile() part. If the master<a name="line.368"></a>
+<span class="sourceLineNo">369</span> * crashes then this function might get called multiple times.<a name="line.369"></a>
+<span class="sourceLineNo">370</span> * <p><a name="line.370"></a>
+<span class="sourceLineNo">371</span> * @param logfile<a name="line.371"></a>
+<span class="sourceLineNo">372</span> * @param conf<a name="line.372"></a>
+<span class="sourceLineNo">373</span> * @throws IOException<a name="line.373"></a>
+<span class="sourceLineNo">374</span> */<a name="line.374"></a>
+<span class="sourceLineNo">375</span> public static void finishSplitLogFile(String logfile,<a name="line.375"></a>
+<span class="sourceLineNo">376</span> Configuration conf) throws IOException {<a name="line.376"></a>
+<span class="sourceLineNo">377</span> Path rootdir = FSUtils.getWALRootDir(conf);<a name="line.377"></a>
+<span class="sourceLineNo">378</span> Path oldLogDir = new Path(rootdir, HConstants.HREGION_OLDLOGDIR_NAME);<a name="line.378"></a>
+<span class="sourceLineNo">379</span> Path logPath;<a name="line.379"></a>
+<span class="sourceLineNo">380</span> if (FSUtils.isStartingWithPath(rootdir, logfile)) {<a name="line.380"></a>
+<span class="sourceLineNo">381</span> logPath = new Path(logfile);<a name="line.381"></a>
+<span class="sourceLineNo">382</span> } else {<a name="line.382"></a>
+<span class="sourceLineNo">383</span> logPath = new Path(rootdir, logfile);<a name="line.383"></a>
+<span class="sourceLineNo">384</span> }<a name="line.384"></a>
+<span class="sourceLineNo">385</span> finishSplitLogFile(rootdir, oldLogDir, logPath, conf);<a name="line.385"></a>
+<span class="sourceLineNo">386</span> }<a name="line.386"></a>
+<span class="sourceLineNo">387</span><a name="line.387"></a>
+<span class="sourceLineNo">388</span> private static void finishSplitLogFile(Path rootdir, Path oldLogDir,<a name="line.388"></a>
+<span class="sourceLineNo">389</span> Path logPath, Configuration conf) throws IOException {<a name="line.389"></a>
+<span class="sourceLineNo">390</span> List<Path> processedLogs = new ArrayList<>();<a name="line.390"></a>
+<span class="sourceLineNo">391</span> List<Path> corruptedLogs = new ArrayList<>();<a name="line.391"></a>
+<span class="sourceLineNo">392</span> FileSystem fs;<a name="line.392"></a>
+<span class="sourceLineNo">393</span> fs = rootdir.getFileSystem(conf);<a name="line.393"></a>
+<span class="sourceLineNo">394</span> if (ZKSplitLog.isCorrupted(rootdir, logPath.getName(), fs)) {<a name="line.394"></a>
+<span class="sourceLineNo">395</span> corruptedLogs.add(logPath);<a name="line.395"></a>
+<span class="sourceLineNo">396</span> } else {<a name="line.396"></a>
+<span class="sourceLineNo">397</span> processedLogs.add(logPath);<a name="line.397"></a>
+<span class="sourceLineNo">398</span> }<a name="line.398"></a>
+<span class="sourceLineNo">399</span> archiveLogs(corruptedLogs, processedLogs, oldLogDir, fs, conf);<a name="line.399"></a>
+<span class="sourceLineNo">400</span> Path stagingDir = ZKSplitLog.getSplitLogDir(rootdir, logPath.getName());<a name="line.400"></a>
+<span class="sourceLineNo">401</span> fs.delete(stagingDir, true);<a name="line.401"></a>
+<span class="sourceLineNo">402</span> }<a name="line.402"></a>
+<span class="sourceLineNo">403</span><a name="line.403"></a>
+<span class="sourceLineNo">404</span> /**<a name="line.404"></a>
+<span class="sourceLineNo">405</span> * Moves processed logs to a oldLogDir after successful processing Moves<a name="line.405"></a>
+<span class="sourceLineNo">406</span> * corrupted logs (any log that couldn't be successfully parsed to corruptDir<a name="line.406"></a>
+<span class="sourceLineNo">407</span> * (.corrupt) for later investigation<a name="line.407"></a>
+<span class="sourceLineNo">408</span> *<a name="line.408"></a>
+<span class="sourceLineNo">409</span> * @param corruptedLogs<a name="line.409"></a>
+<span class="sourceLineNo">410</span> * @param processedLogs<a name="line.410"></a>
+<span class="sourceLineNo">411</span> * @param oldLogDir<a name="line.411"></a>
+<span class="sourceLineNo">412</span> * @param fs<a name="line.412"></a>
+<span class="sourceLineNo">413</span> * @param conf<a name="line.413"></a>
+<span class="sourceLineNo">414</span> * @throws IOException<a name="line.414"></a>
+<span class="sourceLineNo">415</span> */<a name="line.415"></a>
+<span class="sourceLineNo">416</span> private static void archiveLogs(<a name="line.416"></a>
+<span class="sourceLineNo">417</span> final List<Path> corruptedLogs,<a name="line.417"></a>
+<span class="sourceLineNo">418</span> final List<Path> processedLogs, final Path oldLogDir,<a name="line.418"></a>
+<span class="sourceLineNo">419</span> final FileSystem fs, final Configuration conf) throws IOException {<a name="line.419"></a>
+<span class="sourceLineNo">420</span> final Path corruptDir = new Path(FSUtils.getWALRootDir(conf), HConstants.CORRUPT_DIR_NAME);<a name="line.420"></a>
+<span class="sourceLineNo">421</span> if (conf.get("hbase.regionserver.hlog.splitlog.corrupt.dir") != null) {<a name="line.421"></a>
+<span class="sourceLineNo">422</span> LOG.warn("hbase.regionserver.hlog.splitlog.corrupt.dir is deprecated. Default to {}",<a name="line.422"></a>
+<span class="sourceLineNo">423</span> corruptDir);<a name="line.423"></a>
+<span class="sourceLineNo">424</span> }<a name="line.424"></a>
+<span class="sourceLineNo">425</span> if (!fs.mkdirs(corruptDir)) {<a name="line.425"></a>
+<span class="sourceLineNo">426</span> LOG.info("Unable to mkdir {}", corruptDir);<a name="line.426"></a>
+<span class="sourceLineNo">427</span> }<a name="line.427"></a>
+<span class="sourceLineNo">428</span> fs.mkdirs(oldLogDir);<a name="line.428"></a>
+<span class="sourceLineNo">429</span><a name="line.429"></a>
+<span class="sourceLineNo">430</span> // this method can get restarted or called multiple times for archiving<a name="line.430"></a>
+<span class="sourceLineNo">431</span> // the same log files.<a name="line.431"></a>
+<span class="sourceLineNo">432</sp
<TRUNCATED>