http://git-wip-us.apache.org/repos/asf/hbase-site/blob/958717f4/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/WALPlayer.html ---------------------------------------------------------------------- diff --git a/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/WALPlayer.html b/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/WALPlayer.html index 5bed818..8b75278 100644 --- a/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/WALPlayer.html +++ b/apidocs/src-html/org/apache/hadoop/hbase/mapreduce/WALPlayer.html @@ -52,307 +52,314 @@ <span class="sourceLineNo">044</span>import org.apache.hadoop.hbase.client.RegionLocator;<a name="line.44"></a> <span class="sourceLineNo">045</span>import org.apache.hadoop.hbase.client.Table;<a name="line.45"></a> <span class="sourceLineNo">046</span>import org.apache.hadoop.hbase.io.ImmutableBytesWritable;<a name="line.46"></a> -<span class="sourceLineNo">047</span>import org.apache.hadoop.hbase.regionserver.wal.WALEdit;<a name="line.47"></a> -<span class="sourceLineNo">048</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.48"></a> -<span class="sourceLineNo">049</span>import org.apache.hadoop.hbase.wal.WALKey;<a name="line.49"></a> -<span class="sourceLineNo">050</span>import org.apache.hadoop.mapreduce.Job;<a name="line.50"></a> -<span class="sourceLineNo">051</span>import org.apache.hadoop.mapreduce.Mapper;<a name="line.51"></a> -<span class="sourceLineNo">052</span>import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;<a name="line.52"></a> -<span class="sourceLineNo">053</span>import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;<a name="line.53"></a> -<span class="sourceLineNo">054</span>import org.apache.hadoop.util.Tool;<a name="line.54"></a> -<span class="sourceLineNo">055</span>import org.apache.hadoop.util.ToolRunner;<a name="line.55"></a> -<span class="sourceLineNo">056</span><a name="line.56"></a> -<span class="sourceLineNo">057</span>/**<a name="line.57"></a> -<span class="sourceLineNo">058</span> * A tool to replay WAL files as a M/R job.<a name="line.58"></a> -<span class="sourceLineNo">059</span> * The WAL can be replayed for a set of tables or all tables,<a name="line.59"></a> -<span class="sourceLineNo">060</span> * and a timerange can be provided (in milliseconds).<a name="line.60"></a> -<span class="sourceLineNo">061</span> * The WAL is filtered to the passed set of tables and the output<a name="line.61"></a> -<span class="sourceLineNo">062</span> * can optionally be mapped to another set of tables.<a name="line.62"></a> -<span class="sourceLineNo">063</span> *<a name="line.63"></a> -<span class="sourceLineNo">064</span> * WAL replay can also generate HFiles for later bulk importing,<a name="line.64"></a> -<span class="sourceLineNo">065</span> * in that case the WAL is replayed for a single table only.<a name="line.65"></a> -<span class="sourceLineNo">066</span> */<a name="line.66"></a> -<span class="sourceLineNo">067</span>@InterfaceAudience.Public<a name="line.67"></a> -<span class="sourceLineNo">068</span>@InterfaceStability.Stable<a name="line.68"></a> -<span class="sourceLineNo">069</span>public class WALPlayer extends Configured implements Tool {<a name="line.69"></a> -<span class="sourceLineNo">070</span> private static final Log LOG = LogFactory.getLog(WALPlayer.class);<a name="line.70"></a> -<span class="sourceLineNo">071</span> final static String NAME = "WALPlayer";<a name="line.71"></a> -<span class="sourceLineNo">072</span> final static String BULK_OUTPUT_CONF_KEY = "wal.bulk.output";<a name="line.72"></a> -<span class="sourceLineNo">073</span> final static String TABLES_KEY = "wal.input.tables";<a name="line.73"></a> -<span class="sourceLineNo">074</span> final static String TABLE_MAP_KEY = "wal.input.tablesmap";<a name="line.74"></a> -<span class="sourceLineNo">075</span><a name="line.75"></a> -<span class="sourceLineNo">076</span> // This relies on Hadoop Configuration to handle warning about deprecated configs and<a name="line.76"></a> -<span class="sourceLineNo">077</span> // to set the correct non-deprecated configs when an old one shows up.<a name="line.77"></a> -<span class="sourceLineNo">078</span> static {<a name="line.78"></a> -<span class="sourceLineNo">079</span> Configuration.addDeprecation("hlog.bulk.output", BULK_OUTPUT_CONF_KEY);<a name="line.79"></a> -<span class="sourceLineNo">080</span> Configuration.addDeprecation("hlog.input.tables", TABLES_KEY);<a name="line.80"></a> -<span class="sourceLineNo">081</span> Configuration.addDeprecation("hlog.input.tablesmap", TABLE_MAP_KEY);<a name="line.81"></a> -<span class="sourceLineNo">082</span> Configuration.addDeprecation(HLogInputFormat.START_TIME_KEY, WALInputFormat.START_TIME_KEY);<a name="line.82"></a> -<span class="sourceLineNo">083</span> Configuration.addDeprecation(HLogInputFormat.END_TIME_KEY, WALInputFormat.END_TIME_KEY);<a name="line.83"></a> -<span class="sourceLineNo">084</span> }<a name="line.84"></a> -<span class="sourceLineNo">085</span><a name="line.85"></a> -<span class="sourceLineNo">086</span> private final static String JOB_NAME_CONF_KEY = "mapreduce.job.name";<a name="line.86"></a> -<span class="sourceLineNo">087</span><a name="line.87"></a> -<span class="sourceLineNo">088</span> protected WALPlayer(final Configuration c) {<a name="line.88"></a> -<span class="sourceLineNo">089</span> super(c);<a name="line.89"></a> -<span class="sourceLineNo">090</span> }<a name="line.90"></a> -<span class="sourceLineNo">091</span><a name="line.91"></a> -<span class="sourceLineNo">092</span> /**<a name="line.92"></a> -<span class="sourceLineNo">093</span> * A mapper that just writes out KeyValues.<a name="line.93"></a> -<span class="sourceLineNo">094</span> * This one can be used together with {@link KeyValueSortReducer}<a name="line.94"></a> -<span class="sourceLineNo">095</span> */<a name="line.95"></a> -<span class="sourceLineNo">096</span> static class WALKeyValueMapper<a name="line.96"></a> -<span class="sourceLineNo">097</span> extends Mapper<WALKey, WALEdit, ImmutableBytesWritable, KeyValue> {<a name="line.97"></a> -<span class="sourceLineNo">098</span> private byte[] table;<a name="line.98"></a> -<span class="sourceLineNo">099</span><a name="line.99"></a> -<span class="sourceLineNo">100</span> @Override<a name="line.100"></a> -<span class="sourceLineNo">101</span> public void map(WALKey key, WALEdit value,<a name="line.101"></a> -<span class="sourceLineNo">102</span> Context context)<a name="line.102"></a> -<span class="sourceLineNo">103</span> throws IOException {<a name="line.103"></a> -<span class="sourceLineNo">104</span> try {<a name="line.104"></a> -<span class="sourceLineNo">105</span> // skip all other tables<a name="line.105"></a> -<span class="sourceLineNo">106</span> if (Bytes.equals(table, key.getTablename().getName())) {<a name="line.106"></a> -<span class="sourceLineNo">107</span> for (Cell cell : value.getCells()) {<a name="line.107"></a> -<span class="sourceLineNo">108</span> KeyValue kv = KeyValueUtil.ensureKeyValue(cell);<a name="line.108"></a> -<span class="sourceLineNo">109</span> if (WALEdit.isMetaEditFamily(kv)) continue;<a name="line.109"></a> -<span class="sourceLineNo">110</span> context.write(new ImmutableBytesWritable(CellUtil.cloneRow(kv)), kv);<a name="line.110"></a> -<span class="sourceLineNo">111</span> }<a name="line.111"></a> -<span class="sourceLineNo">112</span> }<a name="line.112"></a> -<span class="sourceLineNo">113</span> } catch (InterruptedException e) {<a name="line.113"></a> -<span class="sourceLineNo">114</span> e.printStackTrace();<a name="line.114"></a> -<span class="sourceLineNo">115</span> }<a name="line.115"></a> -<span class="sourceLineNo">116</span> }<a name="line.116"></a> -<span class="sourceLineNo">117</span><a name="line.117"></a> -<span class="sourceLineNo">118</span> @Override<a name="line.118"></a> -<span class="sourceLineNo">119</span> public void setup(Context context) throws IOException {<a name="line.119"></a> -<span class="sourceLineNo">120</span> // only a single table is supported when HFiles are generated with HFileOutputFormat<a name="line.120"></a> -<span class="sourceLineNo">121</span> String[] tables = context.getConfiguration().getStrings(TABLES_KEY);<a name="line.121"></a> -<span class="sourceLineNo">122</span> if (tables == null || tables.length != 1) {<a name="line.122"></a> -<span class="sourceLineNo">123</span> // this can only happen when WALMapper is used directly by a class other than WALPlayer<a name="line.123"></a> -<span class="sourceLineNo">124</span> throw new IOException("Exactly one table must be specified for bulk HFile case.");<a name="line.124"></a> -<span class="sourceLineNo">125</span> }<a name="line.125"></a> -<span class="sourceLineNo">126</span> table = Bytes.toBytes(tables[0]);<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><a name="line.129"></a> -<span class="sourceLineNo">130</span> /**<a name="line.130"></a> -<span class="sourceLineNo">131</span> * A mapper that writes out {@link Mutation} to be directly applied to<a name="line.131"></a> -<span class="sourceLineNo">132</span> * a running HBase instance.<a name="line.132"></a> -<span class="sourceLineNo">133</span> */<a name="line.133"></a> -<span class="sourceLineNo">134</span> protected static class WALMapper<a name="line.134"></a> -<span class="sourceLineNo">135</span> extends Mapper<WALKey, WALEdit, ImmutableBytesWritable, Mutation> {<a name="line.135"></a> -<span class="sourceLineNo">136</span> private Map<TableName, TableName> tables = new TreeMap<TableName, TableName>();<a name="line.136"></a> -<span class="sourceLineNo">137</span><a name="line.137"></a> -<span class="sourceLineNo">138</span> @Override<a name="line.138"></a> -<span class="sourceLineNo">139</span> public void map(WALKey key, WALEdit value, Context context)<a name="line.139"></a> -<span class="sourceLineNo">140</span> throws IOException {<a name="line.140"></a> -<span class="sourceLineNo">141</span> try {<a name="line.141"></a> -<span class="sourceLineNo">142</span> if (tables.isEmpty() || tables.containsKey(key.getTablename())) {<a name="line.142"></a> -<span class="sourceLineNo">143</span> TableName targetTable = tables.isEmpty() ?<a name="line.143"></a> -<span class="sourceLineNo">144</span> key.getTablename() :<a name="line.144"></a> -<span class="sourceLineNo">145</span> tables.get(key.getTablename());<a name="line.145"></a> -<span class="sourceLineNo">146</span> ImmutableBytesWritable tableOut = new ImmutableBytesWritable(targetTable.getName());<a name="line.146"></a> -<span class="sourceLineNo">147</span> Put put = null;<a name="line.147"></a> -<span class="sourceLineNo">148</span> Delete del = null;<a name="line.148"></a> -<span class="sourceLineNo">149</span> Cell lastCell = null;<a name="line.149"></a> -<span class="sourceLineNo">150</span> for (Cell cell : value.getCells()) {<a name="line.150"></a> -<span class="sourceLineNo">151</span> // filtering WAL meta entries<a name="line.151"></a> -<span class="sourceLineNo">152</span> if (WALEdit.isMetaEditFamily(cell)) continue;<a name="line.152"></a> -<span class="sourceLineNo">153</span><a name="line.153"></a> -<span class="sourceLineNo">154</span> // Allow a subclass filter out this cell.<a name="line.154"></a> -<span class="sourceLineNo">155</span> if (filter(context, cell)) {<a name="line.155"></a> -<span class="sourceLineNo">156</span> // A WALEdit may contain multiple operations (HBASE-3584) and/or<a name="line.156"></a> -<span class="sourceLineNo">157</span> // multiple rows (HBASE-5229).<a name="line.157"></a> -<span class="sourceLineNo">158</span> // Aggregate as much as possible into a single Put/Delete<a name="line.158"></a> -<span class="sourceLineNo">159</span> // operation before writing to the context.<a name="line.159"></a> -<span class="sourceLineNo">160</span> if (lastCell == null || lastCell.getTypeByte() != cell.getTypeByte()<a name="line.160"></a> -<span class="sourceLineNo">161</span> || !CellUtil.matchingRow(lastCell, cell)) {<a name="line.161"></a> -<span class="sourceLineNo">162</span> // row or type changed, write out aggregate KVs.<a name="line.162"></a> -<span class="sourceLineNo">163</span> if (put != null) context.write(tableOut, put);<a name="line.163"></a> -<span class="sourceLineNo">164</span> if (del != null) context.write(tableOut, del);<a name="line.164"></a> -<span class="sourceLineNo">165</span> if (CellUtil.isDelete(cell)) {<a name="line.165"></a> -<span class="sourceLineNo">166</span> del = new Delete(CellUtil.cloneRow(cell));<a name="line.166"></a> -<span class="sourceLineNo">167</span> } else {<a name="line.167"></a> -<span class="sourceLineNo">168</span> put = new Put(CellUtil.cloneRow(cell));<a name="line.168"></a> -<span class="sourceLineNo">169</span> }<a name="line.169"></a> -<span class="sourceLineNo">170</span> }<a name="line.170"></a> -<span class="sourceLineNo">171</span> if (CellUtil.isDelete(cell)) {<a name="line.171"></a> -<span class="sourceLineNo">172</span> del.addDeleteMarker(cell);<a name="line.172"></a> -<span class="sourceLineNo">173</span> } else {<a name="line.173"></a> -<span class="sourceLineNo">174</span> put.add(cell);<a name="line.174"></a> -<span class="sourceLineNo">175</span> }<a name="line.175"></a> -<span class="sourceLineNo">176</span> }<a name="line.176"></a> -<span class="sourceLineNo">177</span> lastCell = cell;<a name="line.177"></a> -<span class="sourceLineNo">178</span> }<a name="line.178"></a> -<span class="sourceLineNo">179</span> // write residual KVs<a name="line.179"></a> -<span class="sourceLineNo">180</span> if (put != null) context.write(tableOut, put);<a name="line.180"></a> -<span class="sourceLineNo">181</span> if (del != null) context.write(tableOut, del);<a name="line.181"></a> -<span class="sourceLineNo">182</span> }<a name="line.182"></a> -<span class="sourceLineNo">183</span> } catch (InterruptedException e) {<a name="line.183"></a> -<span class="sourceLineNo">184</span> e.printStackTrace();<a name="line.184"></a> -<span class="sourceLineNo">185</span> }<a name="line.185"></a> -<span class="sourceLineNo">186</span> }<a name="line.186"></a> -<span class="sourceLineNo">187</span><a name="line.187"></a> -<span class="sourceLineNo">188</span> /**<a name="line.188"></a> -<span class="sourceLineNo">189</span> * @param cell<a name="line.189"></a> -<span class="sourceLineNo">190</span> * @return Return true if we are to emit this cell.<a name="line.190"></a> -<span class="sourceLineNo">191</span> */<a name="line.191"></a> -<span class="sourceLineNo">192</span> protected boolean filter(Context context, final Cell cell) {<a name="line.192"></a> -<span class="sourceLineNo">193</span> return true;<a name="line.193"></a> -<span class="sourceLineNo">194</span> }<a name="line.194"></a> -<span class="sourceLineNo">195</span><a name="line.195"></a> -<span class="sourceLineNo">196</span> @Override<a name="line.196"></a> -<span class="sourceLineNo">197</span> public void setup(Context context) throws IOException {<a name="line.197"></a> -<span class="sourceLineNo">198</span> String[] tableMap = context.getConfiguration().getStrings(TABLE_MAP_KEY);<a name="line.198"></a> -<span class="sourceLineNo">199</span> String[] tablesToUse = context.getConfiguration().getStrings(TABLES_KEY);<a name="line.199"></a> -<span class="sourceLineNo">200</span> if (tablesToUse == null && tableMap == null) {<a name="line.200"></a> -<span class="sourceLineNo">201</span> // Then user wants all tables.<a name="line.201"></a> -<span class="sourceLineNo">202</span> } else if (tablesToUse == null || tableMap == null || tablesToUse.length != tableMap.length) {<a name="line.202"></a> -<span class="sourceLineNo">203</span> // this can only happen when WALMapper is used directly by a class other than WALPlayer<a name="line.203"></a> -<span class="sourceLineNo">204</span> throw new IOException("No tables or incorrect table mapping specified.");<a name="line.204"></a> -<span class="sourceLineNo">205</span> }<a name="line.205"></a> -<span class="sourceLineNo">206</span> int i = 0;<a name="line.206"></a> -<span class="sourceLineNo">207</span> if (tablesToUse != null) {<a name="line.207"></a> -<span class="sourceLineNo">208</span> for (String table : tablesToUse) {<a name="line.208"></a> -<span class="sourceLineNo">209</span> tables.put(TableName.valueOf(table),<a name="line.209"></a> -<span class="sourceLineNo">210</span> TableName.valueOf(tableMap[i++]));<a name="line.210"></a> -<span class="sourceLineNo">211</span> }<a name="line.211"></a> -<span class="sourceLineNo">212</span> }<a name="line.212"></a> -<span class="sourceLineNo">213</span> }<a name="line.213"></a> -<span class="sourceLineNo">214</span> }<a name="line.214"></a> -<span class="sourceLineNo">215</span><a name="line.215"></a> -<span class="sourceLineNo">216</span> void setupTime(Configuration conf, String option) throws IOException {<a name="line.216"></a> -<span class="sourceLineNo">217</span> String val = conf.get(option);<a name="line.217"></a> -<span class="sourceLineNo">218</span> if (null == val) return;<a name="line.218"></a> -<span class="sourceLineNo">219</span> long ms;<a name="line.219"></a> -<span class="sourceLineNo">220</span> try {<a name="line.220"></a> -<span class="sourceLineNo">221</span> // first try to parse in user friendly form<a name="line.221"></a> -<span class="sourceLineNo">222</span> ms = new SimpleDateFormat("yyyy-MM-dd'T'HH:mm:ss.SS").parse(val).getTime();<a name="line.222"></a> -<span class="sourceLineNo">223</span> } catch (ParseException pe) {<a name="line.223"></a> -<span class="sourceLineNo">224</span> try {<a name="line.224"></a> -<span class="sourceLineNo">225</span> // then see if just a number of ms's was specified<a name="line.225"></a> -<span class="sourceLineNo">226</span> ms = Long.parseLong(val);<a name="line.226"></a> -<span class="sourceLineNo">227</span> } catch (NumberFormatException nfe) {<a name="line.227"></a> -<span class="sourceLineNo">228</span> throw new IOException(option<a name="line.228"></a> -<span class="sourceLineNo">229</span> + " must be specified either in the form 2001-02-20T16:35:06.99 "<a name="line.229"></a> -<span class="sourceLineNo">230</span> + "or as number of milliseconds");<a name="line.230"></a> -<span class="sourceLineNo">231</span> }<a name="line.231"></a> -<span class="sourceLineNo">232</span> }<a name="line.232"></a> -<span class="sourceLineNo">233</span> conf.setLong(option, ms);<a name="line.233"></a> -<span class="sourceLineNo">234</span> }<a name="line.234"></a> -<span class="sourceLineNo">235</span><a name="line.235"></a> -<span class="sourceLineNo">236</span> /**<a name="line.236"></a> -<span class="sourceLineNo">237</span> * Sets up the actual job.<a name="line.237"></a> -<span class="sourceLineNo">238</span> *<a name="line.238"></a> -<span class="sourceLineNo">239</span> * @param args The command line parameters.<a name="line.239"></a> -<span class="sourceLineNo">240</span> * @return The newly created job.<a name="line.240"></a> -<span class="sourceLineNo">241</span> * @throws IOException When setting up the job fails.<a name="line.241"></a> -<span class="sourceLineNo">242</span> */<a name="line.242"></a> -<span class="sourceLineNo">243</span> public Job createSubmittableJob(String[] args)<a name="line.243"></a> -<span class="sourceLineNo">244</span> throws IOException {<a name="line.244"></a> -<span class="sourceLineNo">245</span> Configuration conf = getConf();<a name="line.245"></a> -<span class="sourceLineNo">246</span> setupTime(conf, HLogInputFormat.START_TIME_KEY);<a name="line.246"></a> -<span class="sourceLineNo">247</span> setupTime(conf, HLogInputFormat.END_TIME_KEY);<a name="line.247"></a> -<span class="sourceLineNo">248</span> Path inputDir = new Path(args[0]);<a name="line.248"></a> -<span class="sourceLineNo">249</span> String[] tables = args[1].split(",");<a name="line.249"></a> -<span class="sourceLineNo">250</span> String[] tableMap;<a name="line.250"></a> -<span class="sourceLineNo">251</span> if (args.length > 2) {<a name="line.251"></a> -<span class="sourceLineNo">252</span> tableMap = args[2].split(",");<a name="line.252"></a> -<span class="sourceLineNo">253</span> if (tableMap.length != tables.length) {<a name="line.253"></a> -<span class="sourceLineNo">254</span> throw new IOException("The same number of tables and mapping must be provided.");<a name="line.254"></a> -<span class="sourceLineNo">255</span> }<a name="line.255"></a> -<span class="sourceLineNo">256</span> } else {<a name="line.256"></a> -<span class="sourceLineNo">257</span> // if not mapping is specified map each table to itself<a name="line.257"></a> -<span class="sourceLineNo">258</span> tableMap = tables;<a name="line.258"></a> -<span class="sourceLineNo">259</span> }<a name="line.259"></a> -<span class="sourceLineNo">260</span> conf.setStrings(TABLES_KEY, tables);<a name="line.260"></a> -<span class="sourceLineNo">261</span> conf.setStrings(TABLE_MAP_KEY, tableMap);<a name="line.261"></a> -<span class="sourceLineNo">262</span> Job job = Job.getInstance(conf, conf.get(JOB_NAME_CONF_KEY, NAME + "_" + inputDir));<a name="line.262"></a> -<span class="sourceLineNo">263</span> job.setJarByClass(WALPlayer.class);<a name="line.263"></a> -<span class="sourceLineNo">264</span> FileInputFormat.setInputPaths(job, inputDir);<a name="line.264"></a> -<span class="sourceLineNo">265</span> job.setInputFormatClass(WALInputFormat.class);<a name="line.265"></a> -<span class="sourceLineNo">266</span> job.setMapOutputKeyClass(ImmutableBytesWritable.class);<a name="line.266"></a> -<span class="sourceLineNo">267</span> String hfileOutPath = conf.get(BULK_OUTPUT_CONF_KEY);<a name="line.267"></a> -<span class="sourceLineNo">268</span> if (hfileOutPath != null) {<a name="line.268"></a> -<span class="sourceLineNo">269</span> // the bulk HFile case<a name="line.269"></a> -<span class="sourceLineNo">270</span> if (tables.length != 1) {<a name="line.270"></a> -<span class="sourceLineNo">271</span> throw new IOException("Exactly one table must be specified for the bulk export option");<a name="line.271"></a> -<span class="sourceLineNo">272</span> }<a name="line.272"></a> -<span class="sourceLineNo">273</span> TableName tableName = TableName.valueOf(tables[0]);<a name="line.273"></a> -<span class="sourceLineNo">274</span> job.setMapperClass(WALKeyValueMapper.class);<a name="line.274"></a> -<span class="sourceLineNo">275</span> job.setReducerClass(KeyValueSortReducer.class);<a name="line.275"></a> -<span class="sourceLineNo">276</span> Path outputDir = new Path(hfileOutPath);<a name="line.276"></a> -<span class="sourceLineNo">277</span> FileOutputFormat.setOutputPath(job, outputDir);<a name="line.277"></a> -<span class="sourceLineNo">278</span> job.setMapOutputValueClass(KeyValue.class);<a name="line.278"></a> -<span class="sourceLineNo">279</span> try (Connection conn = ConnectionFactory.createConnection(conf);<a name="line.279"></a> -<span class="sourceLineNo">280</span> Table table = conn.getTable(tableName);<a name="line.280"></a> -<span class="sourceLineNo">281</span> RegionLocator regionLocator = conn.getRegionLocator(tableName)) {<a name="line.281"></a> -<span class="sourceLineNo">282</span> HFileOutputFormat2.configureIncrementalLoad(job, table.getTableDescriptor(), regionLocator);<a name="line.282"></a> -<span class="sourceLineNo">283</span> }<a name="line.283"></a> -<span class="sourceLineNo">284</span> TableMapReduceUtil.addDependencyJars(job.getConfiguration(),<a name="line.284"></a> -<span class="sourceLineNo">285</span> com.google.common.base.Preconditions.class);<a name="line.285"></a> -<span class="sourceLineNo">286</span> } else {<a name="line.286"></a> -<span class="sourceLineNo">287</span> // output to live cluster<a name="line.287"></a> -<span class="sourceLineNo">288</span> job.setMapperClass(WALMapper.class);<a name="line.288"></a> -<span class="sourceLineNo">289</span> job.setOutputFormatClass(MultiTableOutputFormat.class);<a name="line.289"></a> -<span class="sourceLineNo">290</span> TableMapReduceUtil.addDependencyJars(job);<a name="line.290"></a> -<span class="sourceLineNo">291</span> TableMapReduceUtil.initCredentials(job);<a name="line.291"></a> -<span class="sourceLineNo">292</span> // No reducers.<a name="line.292"></a> -<span class="sourceLineNo">293</span> job.setNumReduceTasks(0);<a name="line.293"></a> -<span class="sourceLineNo">294</span> }<a name="line.294"></a> -<span class="sourceLineNo">295</span> return job;<a name="line.295"></a> -<span class="sourceLineNo">296</span> }<a name="line.296"></a> -<span class="sourceLineNo">297</span><a name="line.297"></a> -<span class="sourceLineNo">298</span> /*<a name="line.298"></a> -<span class="sourceLineNo">299</span> * @param errorMsg Error message. Can be null.<a name="line.299"></a> -<span class="sourceLineNo">300</span> */<a name="line.300"></a> -<span class="sourceLineNo">301</span> private void usage(final String errorMsg) {<a name="line.301"></a> -<span class="sourceLineNo">302</span> if (errorMsg != null && errorMsg.length() > 0) {<a name="line.302"></a> -<span class="sourceLineNo">303</span> System.err.println("ERROR: " + errorMsg);<a name="line.303"></a> -<span class="sourceLineNo">304</span> }<a name="line.304"></a> -<span class="sourceLineNo">305</span> System.err.println("Usage: " + NAME + " [options] <wal inputdir> <tables> [<tableMappings>]");<a name="line.305"></a> -<span class="sourceLineNo">306</span> System.err.println("Read all WAL entries for <tables>.");<a name="line.306"></a> -<span class="sourceLineNo">307</span> System.err.println("If no tables (\"\") are specific, all tables are imported.");<a name="line.307"></a> -<span class="sourceLineNo">308</span> System.err.println("(Careful, even -ROOT- and hbase:meta entries will be imported in that case.)");<a name="line.308"></a> -<span class="sourceLineNo">309</span> System.err.println("Otherwise <tables> is a comma separated list of tables.\n");<a name="line.309"></a> -<span class="sourceLineNo">310</span> System.err.println("The WAL entries can be mapped to new set of tables via <tableMapping>.");<a name="line.310"></a> -<span class="sourceLineNo">311</span> System.err.println("<tableMapping> is a command separated list of targettables.");<a name="line.311"></a> -<span class="sourceLineNo">312</span> System.err.println("If specified, each table in <tables> must have a mapping.\n");<a name="line.312"></a> -<span class="sourceLineNo">313</span> System.err.println("By default " + NAME + " will load data directly into HBase.");<a name="line.313"></a> -<span class="sourceLineNo">314</span> System.err.println("To generate HFiles for a bulk data load instead, pass the option:");<a name="line.314"></a> -<span class="sourceLineNo">315</span> System.err.println(" -D" + BULK_OUTPUT_CONF_KEY + "=/path/for/output");<a name="line.315"></a> -<span class="sourceLineNo">316</span> System.err.println(" (Only one table can be specified, and no mapping is allowed!)");<a name="line.316"></a> -<span class="sourceLineNo">317</span> System.err.println("Other options: (specify time range to WAL edit to consider)");<a name="line.317"></a> -<span class="sourceLineNo">318</span> System.err.println(" -D" + WALInputFormat.START_TIME_KEY + "=[date|ms]");<a name="line.318"></a> -<span class="sourceLineNo">319</span> System.err.println(" -D" + WALInputFormat.END_TIME_KEY + "=[date|ms]");<a name="line.319"></a> -<span class="sourceLineNo">320</span> System.err.println(" -D " + JOB_NAME_CONF_KEY<a name="line.320"></a> -<span class="sourceLineNo">321</span> + "=jobName - use the specified mapreduce job name for the wal player");<a name="line.321"></a> -<span class="sourceLineNo">322</span> System.err.println("For performance also consider the following options:\n"<a name="line.322"></a> -<span class="sourceLineNo">323</span> + " -Dmapreduce.map.speculative=false\n"<a name="line.323"></a> -<span class="sourceLineNo">324</span> + " -Dmapreduce.reduce.speculative=false");<a name="line.324"></a> -<span class="sourceLineNo">325</span> }<a name="line.325"></a> -<span class="sourceLineNo">326</span><a name="line.326"></a> -<span class="sourceLineNo">327</span> /**<a name="line.327"></a> -<span class="sourceLineNo">328</span> * Main entry point.<a name="line.328"></a> -<span class="sourceLineNo">329</span> *<a name="line.329"></a> -<span class="sourceLineNo">330</span> * @param args The command line parameters.<a name="line.330"></a> -<span class="sourceLineNo">331</span> * @throws Exception When running the job fails.<a name="line.331"></a> -<span class="sourceLineNo">332</span> */<a name="line.332"></a> -<span class="sourceLineNo">333</span> public static void main(String[] args) throws Exception {<a name="line.333"></a> -<span class="sourceLineNo">334</span> int ret = ToolRunner.run(new WALPlayer(HBaseConfiguration.create()), args);<a name="line.334"></a> -<span class="sourceLineNo">335</span> System.exit(ret);<a name="line.335"></a> -<span class="sourceLineNo">336</span> }<a name="line.336"></a> -<span class="sourceLineNo">337</span><a name="line.337"></a> -<span class="sourceLineNo">338</span> @Override<a name="line.338"></a> -<span class="sourceLineNo">339</span> public int run(String[] args) throws Exception {<a name="line.339"></a> -<span class="sourceLineNo">340</span> if (args.length < 2) {<a name="line.340"></a> -<span class="sourceLineNo">341</span> usage("Wrong number of arguments: " + args.length);<a name="line.341"></a> -<span class="sourceLineNo">342</span> System.exit(-1);<a name="line.342"></a> -<span class="sourceLineNo">343</span> }<a name="line.343"></a> -<span class="sourceLineNo">344</span> Job job = createSubmittableJob(args);<a name="line.344"></a> -<span class="sourceLineNo">345</span> return job.waitForCompletion(true) ? 0 : 1;<a name="line.345"></a> -<span class="sourceLineNo">346</span> }<a name="line.346"></a> -<span class="sourceLineNo">347</span>}<a name="line.347"></a> +<span class="sourceLineNo">047</span>import org.apache.hadoop.hbase.regionserver.wal.WALCellCodec;<a name="line.47"></a> +<span class="sourceLineNo">048</span>import org.apache.hadoop.hbase.regionserver.wal.WALEdit;<a name="line.48"></a> +<span class="sourceLineNo">049</span>import org.apache.hadoop.hbase.util.Bytes;<a name="line.49"></a> +<span class="sourceLineNo">050</span>import org.apache.hadoop.hbase.wal.WALKey;<a name="line.50"></a> +<span class="sourceLineNo">051</span>import org.apache.hadoop.mapreduce.Job;<a name="line.51"></a> +<span class="sourceLineNo">052</span>import org.apache.hadoop.mapreduce.Mapper;<a name="line.52"></a> +<span class="sourceLineNo">053</span>import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;<a name="line.53"></a> +<span class="sourceLineNo">054</span>import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;<a name="line.54"></a> +<span class="sourceLineNo">055</span>import org.apache.hadoop.util.Tool;<a name="line.55"></a> +<span class="sourceLineNo">056</span>import org.apache.hadoop.util.ToolRunner;<a name="line.56"></a> +<span class="sourceLineNo">057</span><a name="line.57"></a> +<span class="sourceLineNo">058</span>/**<a name="line.58"></a> +<span class="sourceLineNo">059</span> * A tool to replay WAL files as a M/R job.<a name="line.59"></a> +<span class="sourceLineNo">060</span> * The WAL can be replayed for a set of tables or all tables,<a name="line.60"></a> +<span class="sourceLineNo">061</span> * and a timerange can be provided (in milliseconds).<a name="line.61"></a> +<span class="sourceLineNo">062</span> * The WAL is filtered to the passed set of tables and the output<a name="line.62"></a> +<span class="sourceLineNo">063</span> * can optionally be mapped to another set of tables.<a name="line.63"></a> +<span class="sourceLineNo">064</span> *<a name="line.64"></a> +<span class="sourceLineNo">065</span> * WAL replay can also generate HFiles for later bulk importing,<a name="line.65"></a> +<span class="sourceLineNo">066</span> * in that case the WAL is replayed for a single table only.<a name="line.66"></a> +<span class="sourceLineNo">067</span> */<a name="line.67"></a> +<span class="sourceLineNo">068</span>@InterfaceAudience.Public<a name="line.68"></a> +<span class="sourceLineNo">069</span>@InterfaceStability.Stable<a name="line.69"></a> +<span class="sourceLineNo">070</span>public class WALPlayer extends Configured implements Tool {<a name="line.70"></a> +<span class="sourceLineNo">071</span> private static final Log LOG = LogFactory.getLog(WALPlayer.class);<a name="line.71"></a> +<span class="sourceLineNo">072</span> final static String NAME = "WALPlayer";<a name="line.72"></a> +<span class="sourceLineNo">073</span> final static String BULK_OUTPUT_CONF_KEY = "wal.bulk.output";<a name="line.73"></a> +<span class="sourceLineNo">074</span> final static String TABLES_KEY = "wal.input.tables";<a name="line.74"></a> +<span class="sourceLineNo">075</span> final static String TABLE_MAP_KEY = "wal.input.tablesmap";<a name="line.75"></a> +<span class="sourceLineNo">076</span><a name="line.76"></a> +<span class="sourceLineNo">077</span> // This relies on Hadoop Configuration to handle warning about deprecated configs and<a name="line.77"></a> +<span class="sourceLineNo">078</span> // to set the correct non-deprecated configs when an old one shows up.<a name="line.78"></a> +<span class="sourceLineNo">079</span> static {<a name="line.79"></a> +<span class="sourceLineNo">080</span> Configuration.addDeprecation("hlog.bulk.output", BULK_OUTPUT_CONF_KEY);<a name="line.80"></a> +<span class="sourceLineNo">081</span> Configuration.addDeprecation("hlog.input.tables", TABLES_KEY);<a name="line.81"></a> +<span class="sourceLineNo">082</span> Configuration.addDeprecation("hlog.input.tablesmap", TABLE_MAP_KEY);<a name="line.82"></a> +<span class="sourceLineNo">083</span> Configuration.addDeprecation(HLogInputFormat.START_TIME_KEY, WALInputFormat.START_TIME_KEY);<a name="line.83"></a> +<span class="sourceLineNo">084</span> Configuration.addDeprecation(HLogInputFormat.END_TIME_KEY, WALInputFormat.END_TIME_KEY);<a name="line.84"></a> +<span class="sourceLineNo">085</span> }<a name="line.85"></a> +<span class="sourceLineNo">086</span><a name="line.86"></a> +<span class="sourceLineNo">087</span> private final static String JOB_NAME_CONF_KEY = "mapreduce.job.name";<a name="line.87"></a> +<span class="sourceLineNo">088</span><a name="line.88"></a> +<span class="sourceLineNo">089</span> protected WALPlayer(final Configuration c) {<a name="line.89"></a> +<span class="sourceLineNo">090</span> super(c);<a name="line.90"></a> +<span class="sourceLineNo">091</span> }<a name="line.91"></a> +<span class="sourceLineNo">092</span><a name="line.92"></a> +<span class="sourceLineNo">093</span> /**<a name="line.93"></a> +<span class="sourceLineNo">094</span> * A mapper that just writes out KeyValues.<a name="line.94"></a> +<span class="sourceLineNo">095</span> * This one can be used together with {@link KeyValueSortReducer}<a name="line.95"></a> +<span class="sourceLineNo">096</span> */<a name="line.96"></a> +<span class="sourceLineNo">097</span> static class WALKeyValueMapper<a name="line.97"></a> +<span class="sourceLineNo">098</span> extends Mapper<WALKey, WALEdit, ImmutableBytesWritable, KeyValue> {<a name="line.98"></a> +<span class="sourceLineNo">099</span> private byte[] table;<a name="line.99"></a> +<span class="sourceLineNo">100</span><a name="line.100"></a> +<span class="sourceLineNo">101</span> @Override<a name="line.101"></a> +<span class="sourceLineNo">102</span> public void map(WALKey key, WALEdit value,<a name="line.102"></a> +<span class="sourceLineNo">103</span> Context context)<a name="line.103"></a> +<span class="sourceLineNo">104</span> throws IOException {<a name="line.104"></a> +<span class="sourceLineNo">105</span> try {<a name="line.105"></a> +<span class="sourceLineNo">106</span> // skip all other tables<a name="line.106"></a> +<span class="sourceLineNo">107</span> if (Bytes.equals(table, key.getTablename().getName())) {<a name="line.107"></a> +<span class="sourceLineNo">108</span> for (Cell cell : value.getCells()) {<a name="line.108"></a> +<span class="sourceLineNo">109</span> KeyValue kv = KeyValueUtil.ensureKeyValue(cell);<a name="line.109"></a> +<span class="sourceLineNo">110</span> if (WALEdit.isMetaEditFamily(kv)) continue;<a name="line.110"></a> +<span class="sourceLineNo">111</span> context.write(new ImmutableBytesWritable(CellUtil.cloneRow(kv)), kv);<a name="line.111"></a> +<span class="sourceLineNo">112</span> }<a name="line.112"></a> +<span class="sourceLineNo">113</span> }<a name="line.113"></a> +<span class="sourceLineNo">114</span> } catch (InterruptedException e) {<a name="line.114"></a> +<span class="sourceLineNo">115</span> e.printStackTrace();<a name="line.115"></a> +<span class="sourceLineNo">116</span> }<a name="line.116"></a> +<span class="sourceLineNo">117</span> }<a name="line.117"></a> +<span class="sourceLineNo">118</span><a name="line.118"></a> +<span class="sourceLineNo">119</span> @Override<a name="line.119"></a> +<span class="sourceLineNo">120</span> public void setup(Context context) throws IOException {<a name="line.120"></a> +<span class="sourceLineNo">121</span> // only a single table is supported when HFiles are generated with HFileOutputFormat<a name="line.121"></a> +<span class="sourceLineNo">122</span> String[] tables = context.getConfiguration().getStrings(TABLES_KEY);<a name="line.122"></a> +<span class="sourceLineNo">123</span> if (tables == null || tables.length != 1) {<a name="line.123"></a> +<span class="sourceLineNo">124</span> // this can only happen when WALMapper is used directly by a class other than WALPlayer<a name="line.124"></a> +<span class="sourceLineNo">125</span> throw new IOException("Exactly one table must be specified for bulk HFile case.");<a name="line.125"></a> +<span class="sourceLineNo">126</span> }<a name="line.126"></a> +<span class="sourceLineNo">127</span> table = Bytes.toBytes(tables[0]);<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><a name="line.130"></a> +<span class="sourceLineNo">131</span> /**<a name="line.131"></a> +<span class="sourceLineNo">132</span> * A mapper that writes out {@link Mutation} to be directly applied to<a name="line.132"></a> +<span class="sourceLineNo">133</span> * a running HBase instance.<a name="line.133"></a> +<span class="sourceLineNo">134</span> */<a name="line.134"></a> +<span class="sourceLineNo">135</span> protected static class WALMapper<a name="line.135"></a> +<span class="sourceLineNo">136</span> extends Mapper<WALKey, WALEdit, ImmutableBytesWritable, Mutation> {<a name="line.136"></a> +<span class="sourceLineNo">137</span> private Map<TableName, TableName> tables = new TreeMap<TableName, TableName>();<a name="line.137"></a> +<span class="sourceLineNo">138</span><a name="line.138"></a> +<span class="sourceLineNo">139</span> @Override<a name="line.139"></a> +<span class="sourceLineNo">140</span> public void map(WALKey key, WALEdit value, Context context)<a name="line.140"></a> +<span class="sourceLineNo">141</span> throws IOException {<a name="line.141"></a> +<span class="sourceLineNo">142</span> try {<a name="line.142"></a> +<span class="sourceLineNo">143</span> if (tables.isEmpty() || tables.containsKey(key.getTablename())) {<a name="line.143"></a> +<span class="sourceLineNo">144</span> TableName targetTable = tables.isEmpty() ?<a name="line.144"></a> +<span class="sourceLineNo">145</span> key.getTablename() :<a name="line.145"></a> +<span class="sourceLineNo">146</span> tables.get(key.getTablename());<a name="line.146"></a> +<span class="sourceLineNo">147</span> ImmutableBytesWritable tableOut = new ImmutableBytesWritable(targetTable.getName());<a name="line.147"></a> +<span class="sourceLineNo">148</span> Put put = null;<a name="line.148"></a> +<span class="sourceLineNo">149</span> Delete del = null;<a name="line.149"></a> +<span class="sourceLineNo">150</span> Cell lastCell = null;<a name="line.150"></a> +<span class="sourceLineNo">151</span> for (Cell cell : value.getCells()) {<a name="line.151"></a> +<span class="sourceLineNo">152</span> // filtering WAL meta entries<a name="line.152"></a> +<span class="sourceLineNo">153</span> if (WALEdit.isMetaEditFamily(cell)) continue;<a name="line.153"></a> +<span class="sourceLineNo">154</span><a name="line.154"></a> +<span class="sourceLineNo">155</span> // Allow a subclass filter out this cell.<a name="line.155"></a> +<span class="sourceLineNo">156</span> if (filter(context, cell)) {<a name="line.156"></a> +<span class="sourceLineNo">157</span> // A WALEdit may contain multiple operations (HBASE-3584) and/or<a name="line.157"></a> +<span class="sourceLineNo">158</span> // multiple rows (HBASE-5229).<a name="line.158"></a> +<span class="sourceLineNo">159</span> // Aggregate as much as possible into a single Put/Delete<a name="line.159"></a> +<span class="sourceLineNo">160</span> // operation before writing to the context.<a name="line.160"></a> +<span class="sourceLineNo">161</span> if (lastCell == null || lastCell.getTypeByte() != cell.getTypeByte()<a name="line.161"></a> +<span class="sourceLineNo">162</span> || !CellUtil.matchingRow(lastCell, cell)) {<a name="line.162"></a> +<span class="sourceLineNo">163</span> // row or type changed, write out aggregate KVs.<a name="line.163"></a> +<span class="sourceLineNo">164</span> if (put != null) context.write(tableOut, put);<a name="line.164"></a> +<span class="sourceLineNo">165</span> if (del != null) context.write(tableOut, del);<a name="line.165"></a> +<span class="sourceLineNo">166</span> if (CellUtil.isDelete(cell)) {<a name="line.166"></a> +<span class="sourceLineNo">167</span> del = new Delete(CellUtil.cloneRow(cell));<a name="line.167"></a> +<span class="sourceLineNo">168</span> } else {<a name="line.168"></a> +<span class="sourceLineNo">169</span> put = new Put(CellUtil.cloneRow(cell));<a name="line.169"></a> +<span class="sourceLineNo">170</span> }<a name="line.170"></a> +<span class="sourceLineNo">171</span> }<a name="line.171"></a> +<span class="sourceLineNo">172</span> if (CellUtil.isDelete(cell)) {<a name="line.172"></a> +<span class="sourceLineNo">173</span> del.addDeleteMarker(cell);<a name="line.173"></a> +<span class="sourceLineNo">174</span> } else {<a name="line.174"></a> +<span class="sourceLineNo">175</span> put.add(cell);<a name="line.175"></a> +<span class="sourceLineNo">176</span> }<a name="line.176"></a> +<span class="sourceLineNo">177</span> }<a name="line.177"></a> +<span class="sourceLineNo">178</span> lastCell = cell;<a name="line.178"></a> +<span class="sourceLineNo">179</span> }<a name="line.179"></a> +<span class="sourceLineNo">180</span> // write residual KVs<a name="line.180"></a> +<span class="sourceLineNo">181</span> if (put != null) context.write(tableOut, put);<a name="line.181"></a> +<span class="sourceLineNo">182</span> if (del != null) context.write(tableOut, del);<a name="line.182"></a> +<span class="sourceLineNo">183</span> }<a name="line.183"></a> +<span class="sourceLineNo">184</span> } catch (InterruptedException e) {<a name="line.184"></a> +<span class="sourceLineNo">185</span> e.printStackTrace();<a name="line.185"></a> +<span class="sourceLineNo">186</span> }<a name="line.186"></a> +<span class="sourceLineNo">187</span> }<a name="line.187"></a> +<span class="sourceLineNo">188</span><a name="line.188"></a> +<span class="sourceLineNo">189</span> /**<a name="line.189"></a> +<span class="sourceLineNo">190</span> * @param cell<a name="line.190"></a> +<span class="sourceLineNo">191</span> * @return Return true if we are to emit this cell.<a name="line.191"></a> +<span class="sourceLineNo">192</span> */<a name="line.192"></a> +<span class="sourceLineNo">193</span> protected boolean filter(Context context, final Cell cell) {<a name="line.193"></a> +<span class="sourceLineNo">194</span> return true;<a name="line.194"></a> +<span class="sourceLineNo">195</span> }<a name="line.195"></a> +<span class="sourceLineNo">196</span><a name="line.196"></a> +<span class="sourceLineNo">197</span> @Override<a name="line.197"></a> +<span class="sourceLineNo">198</span> public void setup(Context context) throws IOException {<a name="line.198"></a> +<span class="sourceLineNo">199</span> String[] tableMap = context.getConfiguration().getStrings(TABLE_MAP_KEY);<a name="line.199"></a> +<span class="sourceLineNo">200</span> String[] tablesToUse = context.getConfiguration().getStrings(TABLES_KEY);<a name="line.200"></a> +<span class="sourceLineNo">201</span> if (tablesToUse == null && tableMap == null) {<a name="line.201"></a> +<span class="sourceLineNo">202</span> // Then user wants all tables.<a name="line.202"></a> +<span class="sourceLineNo">203</span> } else if (tablesToUse == null || tableMap == null || tablesToUse.length != tableMap.length) {<a name="line.203"></a> +<span class="sourceLineNo">204</span> // this can only happen when WALMapper is used directly by a class other than WALPlayer<a name="line.204"></a> +<span class="sourceLineNo">205</span> throw new IOException("No tables or incorrect table mapping specified.");<a name="line.205"></a> +<span class="sourceLineNo">206</span> }<a name="line.206"></a> +<span class="sourceLineNo">207</span> int i = 0;<a name="line.207"></a> +<span class="sourceLineNo">208</span> if (tablesToUse != null) {<a name="line.208"></a> +<span class="sourceLineNo">209</span> for (String table : tablesToUse) {<a name="line.209"></a> +<span class="sourceLineNo">210</span> tables.put(TableName.valueOf(table),<a name="line.210"></a> +<span class="sourceLineNo">211</span> TableName.valueOf(tableMap[i++]));<a name="line.211"></a> +<span class="sourceLineNo">212</span> }<a name="line.212"></a> +<span class="sourceLineNo">213</span> }<a name="line.213"></a> +<span class="sourceLineNo">214</span> }<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> void setupTime(Configuration conf, String option) throws IOException {<a name="line.217"></a> +<span class="sourceLineNo">218</span> String val = conf.get(option);<a name="line.218"></a> +<span class="sourceLineNo">219</span> if (null == val) return;<a name="line.219"></a> +<span class="sourceLineNo">220</span> long ms;<a name="line.220"></a> +<span class="sourceLineNo">221</span> try {<a name="line.221"></a> +<span class="sourceLineNo">222</span> // first try to parse in user friendly form<a name="line.222"></a> +<span class="sourceLineNo">223</span> ms = new SimpleDateFormat("yyyy-MM-dd'T'HH:mm:ss.SS").parse(val).getTime();<a name="line.223"></a> +<span class="sourceLineNo">224</span> } catch (ParseException pe) {<a name="line.224"></a> +<span class="sourceLineNo">225</span> try {<a name="line.225"></a> +<span class="sourceLineNo">226</span> // then see if just a number of ms's was specified<a name="line.226"></a> +<span class="sourceLineNo">227</span> ms = Long.parseLong(val);<a name="line.227"></a> +<span class="sourceLineNo">228</span> } catch (NumberFormatException nfe) {<a name="line.228"></a> +<span class="sourceLineNo">229</span> throw new IOException(option<a name="line.229"></a> +<span class="sourceLineNo">230</span> + " must be specified either in the form 2001-02-20T16:35:06.99 "<a name="line.230"></a> +<span class="sourceLineNo">231</span> + "or as number of milliseconds");<a name="line.231"></a> +<span class="sourceLineNo">232</span> }<a name="line.232"></a> +<span class="sourceLineNo">233</span> }<a name="line.233"></a> +<span class="sourceLineNo">234</span> conf.setLong(option, ms);<a name="line.234"></a> +<span class="sourceLineNo">235</span> }<a name="line.235"></a> +<span class="sourceLineNo">236</span><a name="line.236"></a> +<span class="sourceLineNo">237</span> /**<a name="line.237"></a> +<span class="sourceLineNo">238</span> * Sets up the actual job.<a name="line.238"></a> +<span class="sourceLineNo">239</span> *<a name="line.239"></a> +<span class="sourceLineNo">240</span> * @param args The command line parameters.<a name="line.240"></a> +<span class="sourceLineNo">241</span> * @return The newly created job.<a name="line.241"></a> +<span class="sourceLineNo">242</span> * @throws IOException When setting up the job fails.<a name="line.242"></a> +<span class="sourceLineNo">243</span> */<a name="line.243"></a> +<span class="sourceLineNo">244</span> public Job createSubmittableJob(String[] args)<a name="line.244"></a> +<span class="sourceLineNo">245</span> throws IOException {<a name="line.245"></a> +<span class="sourceLineNo">246</span> Configuration conf = getConf();<a name="line.246"></a> +<span class="sourceLineNo">247</span> setupTime(conf, HLogInputFormat.START_TIME_KEY);<a name="line.247"></a> +<span class="sourceLineNo">248</span> setupTime(conf, HLogInputFormat.END_TIME_KEY);<a name="line.248"></a> +<span class="sourceLineNo">249</span> Path inputDir = new Path(args[0]);<a name="line.249"></a> +<span class="sourceLineNo">250</span> String[] tables = args[1].split(",");<a name="line.250"></a> +<span class="sourceLineNo">251</span> String[] tableMap;<a name="line.251"></a> +<span class="sourceLineNo">252</span> if (args.length > 2) {<a name="line.252"></a> +<span class="sourceLineNo">253</span> tableMap = args[2].split(",");<a name="line.253"></a> +<span class="sourceLineNo">254</span> if (tableMap.length != tables.length) {<a name="line.254"></a> +<span class="sourceLineNo">255</span> throw new IOException("The same number of tables and mapping must be provided.");<a name="line.255"></a> +<span class="sourceLineNo">256</span> }<a name="line.256"></a> +<span class="sourceLineNo">257</span> } else {<a name="line.257"></a> +<span class="sourceLineNo">258</span> // if not mapping is specified map each table to itself<a name="line.258"></a> +<span class="sourceLineNo">259</span> tableMap = tables;<a name="line.259"></a> +<span class="sourceLineNo">260</span> }<a name="line.260"></a> +<span class="sourceLineNo">261</span> conf.setStrings(TABLES_KEY, tables);<a name="line.261"></a> +<span class="sourceLineNo">262</span> conf.setStrings(TABLE_MAP_KEY, tableMap);<a name="line.262"></a> +<span class="sourceLineNo">263</span> Job job = Job.getInstance(conf, conf.get(JOB_NAME_CONF_KEY, NAME + "_" + inputDir));<a name="line.263"></a> +<span class="sourceLineNo">264</span> job.setJarByClass(WALPlayer.class);<a name="line.264"></a> +<span class="sourceLineNo">265</span> FileInputFormat.setInputPaths(job, inputDir);<a name="line.265"></a> +<span class="sourceLineNo">266</span> job.setInputFormatClass(WALInputFormat.class);<a name="line.266"></a> +<span class="sourceLineNo">267</span> job.setMapOutputKeyClass(ImmutableBytesWritable.class);<a name="line.267"></a> +<span class="sourceLineNo">268</span> String hfileOutPath = conf.get(BULK_OUTPUT_CONF_KEY);<a name="line.268"></a> +<span class="sourceLineNo">269</span> if (hfileOutPath != null) {<a name="line.269"></a> +<span class="sourceLineNo">270</span> // the bulk HFile case<a name="line.270"></a> +<span class="sourceLineNo">271</span> if (tables.length != 1) {<a name="line.271"></a> +<span class="sourceLineNo">272</span> throw new IOException("Exactly one table must be specified for the bulk export option");<a name="line.272"></a> +<span class="sourceLineNo">273</span> }<a name="line.273"></a> +<span class="sourceLineNo">274</span> TableName tableName = TableName.valueOf(tables[0]);<a name="line.274"></a> +<span class="sourceLineNo">275</span> job.setMapperClass(WALKeyValueMapper.class);<a name="line.275"></a> +<span class="sourceLineNo">276</span> job.setReducerClass(KeyValueSortReducer.class);<a name="line.276"></a> +<span class="sourceLineNo">277</span> Path outputDir = new Path(hfileOutPath);<a name="line.277"></a> +<span class="sourceLineNo">278</span> FileOutputFormat.setOutputPath(job, outputDir);<a name="line.278"></a> +<span class="sourceLineNo">279</span> job.setMapOutputValueClass(KeyValue.class);<a name="line.279"></a> +<span class="sourceLineNo">280</span> try (Connection conn = ConnectionFactory.createConnection(conf);<a name="line.280"></a> +<span class="sourceLineNo">281</span> Table table = conn.getTable(tableName);<a name="line.281"></a> +<span class="sourceLineNo">282</span> RegionLocator regionLocator = conn.getRegionLocator(tableName)) {<a name="line.282"></a> +<span class="sourceLineNo">283</span> HFileOutputFormat2.configureIncrementalLoad(job, table.getTableDescriptor(), regionLocator);<a name="line.283"></a> +<span class="sourceLineNo">284</span> }<a name="line.284"></a> +<span class="sourceLineNo">285</span> TableMapReduceUtil.addDependencyJars(job.getConfiguration(),<a name="line.285"></a> +<span class="sourceLineNo">286</span> com.google.common.base.Preconditions.class);<a name="line.286"></a> +<span class="sourceLineNo">287</span> } else {<a name="line.287"></a> +<span class="sourceLineNo">288</span> // output to live cluster<a name="line.288"></a> +<span class="sourceLineNo">289</span> job.setMapperClass(WALMapper.class);<a name="line.289"></a> +<span class="sourceLineNo">290</span> job.setOutputFormatClass(MultiTableOutputFormat.class);<a name="line.290"></a> +<span class="sourceLineNo">291</span> TableMapReduceUtil.addDependencyJars(job);<a name="line.291"></a> +<span class="sourceLineNo">292</span> TableMapReduceUtil.initCredentials(job);<a name="line.292"></a> +<span class="sourceLineNo">293</span> // No reducers.<a name="line.293"></a> +<span class="sourceLineNo">294</span> job.setNumReduceTasks(0);<a name="line.294"></a> +<span class="sourceLineNo">295</span> }<a name="line.295"></a> +<span class="sourceLineNo">296</span> String codecCls = WALCellCodec.getWALCellCodecClass(conf);<a name="line.296"></a> +<span class="sourceLineNo">297</span> try {<a name="line.297"></a> +<span class="sourceLineNo">298</span> TableMapReduceUtil.addDependencyJars(conf, Class.forName(codecCls));<a name="line.298"></a> +<span class="sourceLineNo">299</span> } catch (Exception e) {<a name="line.299"></a> +<span class="sourceLineNo">300</span> throw new IOException("Cannot determine wal codec class " + codecCls, e);<a name="line.300"></a> +<span class="sourceLineNo">301</span> }<a name="line.301"></a> +<span class="sourceLineNo">302</span> return job;<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> /*<a name="line.305"></a> +<span class="sourceLineNo">306</span> * @param errorMsg Error message. Can be null.<a name="line.306"></a> +<span class="sourceLineNo">307</span> */<a name="line.307"></a> +<span class="sourceLineNo">308</span> private void usage(final String errorMsg) {<a name="line.308"></a> +<span class="sourceLineNo">309</span> if (errorMsg != null && errorMsg.length() > 0) {<a name="line.309"></a> +<span class="sourceLineNo">310</span> System.err.println("ERROR: " + errorMsg);<a name="line.310"></a> +<span class="sourceLineNo">311</span> }<a name="line.311"></a> +<span class="sourceLineNo">312</span> System.err.println("Usage: " + NAME + " [options] <wal inputdir> <tables> [<tableMappings>]");<a name="line.312"></a> +<span class="sourceLineNo">313</span> System.err.println("Read all WAL entries for <tables>.");<a name="line.313"></a> +<span class="sourceLineNo">314</span> System.err.println("If no tables (\"\") are specific, all tables are imported.");<a name="line.314"></a> +<span class="sourceLineNo">315</span> System.err.println("(Careful, even -ROOT- and hbase:meta entries will be imported in that case.)");<a name="line.315"></a> +<span class="sourceLineNo">316</span> System.err.println("Otherwise <tables> is a comma separated list of tables.\n");<a name="line.316"></a> +<span class="sourceLineNo">317</span> System.err.println("The WAL entries can be mapped to new set of tables via <tableMapping>.");<a name="line.317"></a> +<span class="sourceLineNo">318</span> System.err.println("<tableMapping> is a command separated list of targettables.");<a name="line.318"></a> +<span class="sourceLineNo">319</span> System.err.println("If specified, each table in <tables> must have a mapping.\n");<a name="line.319"></a> +<span class="sourceLineNo">320</span> System.err.println("By default " + NAME + " will load data directly into HBase.");<a name="line.320"></a> +<span class="sourceLineNo">321</span> System.err.println("To generate HFiles for a bulk data load instead, pass the option:");<a name="line.321"></a> +<span class="sourceLineNo">322</span> System.err.println(" -D" + BULK_OUTPUT_CONF_KEY + "=/path/for/output");<a name="line.322"></a> +<span class="sourceLineNo">323</span> System.err.println(" (Only one table can be specified, and no mapping is allowed!)");<a name="line.323"></a> +<span class="sourceLineNo">324</span> System.err.println("Other options: (specify time range to WAL edit to consider)");<a name="line.324"></a> +<span class="sourceLineNo">325</span> System.err.println(" -D" + WALInputFormat.START_TIME_KEY + "=[date|ms]");<a name="line.325"></a> +<span class="sourceLineNo">326</span> System.err.println(" -D" + WALInputFormat.END_TIME_KEY + "=[date|ms]");<a name="line.326"></a> +<span class="sourceLineNo">327</span> System.err.println(" -D " + JOB_NAME_CONF_KEY<a name="line.327"></a> +<span class="sourceLineNo">328</span> + "=jobName - use the specified mapreduce job name for the wal player");<a name="line.328"></a> +<span class="sourceLineNo">329</span> System.err.println("For performance also consider the following options:\n"<a name="line.329"></a> +<span class="sourceLineNo">330</span> + " -Dmapreduce.map.speculative=false\n"<a name="line.330"></a> +<span class="sourceLineNo">331</span> + " -Dmapreduce.reduce.speculative=false");<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> /**<a name="line.334"></a> +<span class="sourceLineNo">335</span> * Main entry point.<a name="line.335"></a> +<span class="sourceLineNo">336</span> *<a name="line.336"></a> +<span class="sourceLineNo">337</span> * @param args The command line parameters.<a name="line.337"></a> +<span class="sourceLineNo">338</span> * @throws Exception When running the job fails.<a name="line.338"></a> +<span class="sourceLineNo">339</span> */<a name="line.339"></a> +<span class="sourceLineNo">340</span> public static void main(String[] args) throws Exception {<a name="line.340"></a> +<span class="sourceLineNo">341</span> int ret = ToolRunner.run(new WALPlayer(HBaseConfiguration.create()), args);<a name="line.341"></a> +<span class="sourceLineNo">342</span> System.exit(ret);<a name="line.342"></a> +<span class="sourceLineNo">343</span> }<a name="line.343"></a> +<span class="sourceLineNo">344</span><a name="line.344"></a> +<span class="sourceLineNo">345</span> @Override<a name="line.345"></a> +<span class="sourceLineNo">346</span> public int run(String[] args) throws Exception {<a name="line.346"></a> +<span class="sourceLineNo">347</span> if (args.length < 2) {<a name="line.347"></a> +<span class="sourceLineNo">348</span> usage("Wrong number of arguments: " + args.length);<a name="line.348"></a> +<span class="sourceLineNo">349</span> System.exit(-1);<a name="line.349"></a> +<span class="sourceLineNo">350</span> }<a name="line.350"></a> +<span class="sourceLineNo">351</span> Job job = createSubmittableJob(args);<a name="line.351"></a> +<span class="sourceLineNo">352</span> return job.waitForCompletion(true) ? 0 : 1;<a name="line.352"></a> +<span class="sourceLineNo">353</span> }<a name="line.353"></a> +<span class="sourceLineNo">354</span>}<a name="line.354"></a>
http://git-wip-us.apache.org/repos/asf/hbase-site/blob/958717f4/apidocs/src-html/org/apache/hadoop/hbase/snapshot/CorruptedSnapshotException.html ---------------------------------------------------------------------- diff --git a/apidocs/src-html/org/apache/hadoop/hbase/snapshot/CorruptedSnapshotException.html b/apidocs/src-html/org/apache/hadoop/hbase/snapshot/CorruptedSnapshotException.html index 1bdac62..c260402 100644 --- a/apidocs/src-html/org/apache/hadoop/hbase/snapshot/CorruptedSnapshotException.html +++ b/apidocs/src-html/org/apache/hadoop/hbase/snapshot/CorruptedSnapshotException.html @@ -27,7 +27,7 @@ <span class="sourceLineNo">019</span><a name="line.19"></a> <span class="sourceLineNo">020</span>import org.apache.hadoop.hbase.classification.InterfaceAudience;<a name="line.20"></a> <span class="sourceLineNo">021</span>import org.apache.hadoop.hbase.classification.InterfaceStability;<a name="line.21"></a> -<span class="sourceLineNo">022</span>import org.apache.hadoop.hbase.protobuf.generated.HBaseProtos.SnapshotDescription;<a name="line.22"></a> +<span class="sourceLineNo">022</span>import org.apache.hadoop.hbase.client.SnapshotDescription;<a name="line.22"></a> <span class="sourceLineNo">023</span><a name="line.23"></a> <span class="sourceLineNo">024</span><a name="line.24"></a> <span class="sourceLineNo">025</span>/**<a name="line.25"></a> http://git-wip-us.apache.org/repos/asf/hbase-site/blob/958717f4/apidocs/src-html/org/apache/hadoop/hbase/snapshot/HBaseSnapshotException.html ---------------------------------------------------------------------- diff --git a/apidocs/src-html/org/apache/hadoop/hbase/snapshot/HBaseSnapshotException.html b/apidocs/src-html/org/apache/hadoop/hbase/snapshot/HBaseSnapshotException.html index 13009e9..9ea5bdd 100644 --- a/apidocs/src-html/org/apache/hadoop/hbase/snapshot/HBaseSnapshotException.html +++ b/apidocs/src-html/org/apache/hadoop/hbase/snapshot/HBaseSnapshotException.html @@ -28,7 +28,7 @@ <span class="sourceLineNo">020</span>import org.apache.hadoop.hbase.DoNotRetryIOException;<a name="line.20"></a> <span class="sourceLineNo">021</span>import org.apache.hadoop.hbase.classification.InterfaceAudience;<a name="line.21"></a> <span class="sourceLineNo">022</span>import org.apache.hadoop.hbase.classification.InterfaceStability;<a name="line.22"></a> -<span class="sourceLineNo">023</span>import org.apache.hadoop.hbase.protobuf.generated.HBaseProtos.SnapshotDescription;<a name="line.23"></a> +<span class="sourceLineNo">023</span>import org.apache.hadoop.hbase.client.SnapshotDescription;<a name="line.23"></a> <span class="sourceLineNo">024</span><a name="line.24"></a> <span class="sourceLineNo">025</span>/**<a name="line.25"></a> <span class="sourceLineNo">026</span> * General exception base class for when a snapshot fails<a name="line.26"></a> http://git-wip-us.apache.org/repos/asf/hbase-site/blob/958717f4/apidocs/src-html/org/apache/hadoop/hbase/snapshot/RestoreSnapshotException.html ---------------------------------------------------------------------- diff --git a/apidocs/src-html/org/apache/hadoop/hbase/snapshot/RestoreSnapshotException.html b/apidocs/src-html/org/apache/hadoop/hbase/snapshot/RestoreSnapshotException.html index bc641a1..e7c132e 100644 --- a/apidocs/src-html/org/apache/hadoop/hbase/snapshot/RestoreSnapshotException.html +++ b/apidocs/src-html/org/apache/hadoop/hbase/snapshot/RestoreSnapshotException.html @@ -28,7 +28,7 @@ <span class="sourceLineNo">020</span><a name="line.20"></a> <span class="sourceLineNo">021</span>import org.apache.hadoop.hbase.classification.InterfaceAudience;<a name="line.21"></a> <span class="sourceLineNo">022</span>import org.apache.hadoop.hbase.classification.InterfaceStability;<a name="line.22"></a> -<span class="sourceLineNo">023</span>import org.apache.hadoop.hbase.protobuf.generated.HBaseProtos.SnapshotDescription;<a name="line.23"></a> +<span class="sourceLineNo">023</span>import org.apache.hadoop.hbase.client.SnapshotDescription;<a name="line.23"></a> <span class="sourceLineNo">024</span><a name="line.24"></a> <span class="sourceLineNo">025</span>/**<a name="line.25"></a> <span class="sourceLineNo">026</span> * Thrown when a snapshot could not be restored due to a server-side error when restoring it.<a name="line.26"></a> http://git-wip-us.apache.org/repos/asf/hbase-site/blob/958717f4/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotCreationException.html ---------------------------------------------------------------------- diff --git a/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotCreationException.html b/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotCreationException.html index b060a0a..6f0f77f 100644 --- a/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotCreationException.html +++ b/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotCreationException.html @@ -27,7 +27,7 @@ <span class="sourceLineNo">019</span><a name="line.19"></a> <span class="sourceLineNo">020</span>import org.apache.hadoop.hbase.classification.InterfaceAudience;<a name="line.20"></a> <span class="sourceLineNo">021</span>import org.apache.hadoop.hbase.classification.InterfaceStability;<a name="line.21"></a> -<span class="sourceLineNo">022</span>import org.apache.hadoop.hbase.protobuf.generated.HBaseProtos.SnapshotDescription;<a name="line.22"></a> +<span class="sourceLineNo">022</span>import org.apache.hadoop.hbase.client.SnapshotDescription;<a name="line.22"></a> <span class="sourceLineNo">023</span><a name="line.23"></a> <span class="sourceLineNo">024</span>/**<a name="line.24"></a> <span class="sourceLineNo">025</span> * Thrown when a snapshot could not be created due to a server-side error when<a name="line.25"></a> http://git-wip-us.apache.org/repos/asf/hbase-site/blob/958717f4/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotDoesNotExistException.html ---------------------------------------------------------------------- diff --git a/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotDoesNotExistException.html b/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotDoesNotExistException.html index d5d61e0..b63a983 100644 --- a/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotDoesNotExistException.html +++ b/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotDoesNotExistException.html @@ -27,7 +27,7 @@ <span class="sourceLineNo">019</span><a name="line.19"></a> <span class="sourceLineNo">020</span>import org.apache.hadoop.hbase.classification.InterfaceAudience;<a name="line.20"></a> <span class="sourceLineNo">021</span>import org.apache.hadoop.hbase.classification.InterfaceStability;<a name="line.21"></a> -<span class="sourceLineNo">022</span>import org.apache.hadoop.hbase.protobuf.generated.HBaseProtos.SnapshotDescription;<a name="line.22"></a> +<span class="sourceLineNo">022</span>import org.apache.hadoop.hbase.client.SnapshotDescription;<a name="line.22"></a> <span class="sourceLineNo">023</span><a name="line.23"></a> <span class="sourceLineNo">024</span><a name="line.24"></a> <span class="sourceLineNo">025</span>/**<a name="line.25"></a> http://git-wip-us.apache.org/repos/asf/hbase-site/blob/958717f4/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotExistsException.html ---------------------------------------------------------------------- diff --git a/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotExistsException.html b/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotExistsException.html index cd4e771..ef29435 100644 --- a/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotExistsException.html +++ b/apidocs/src-html/org/apache/hadoop/hbase/snapshot/SnapshotExistsException.html @@ -27,7 +27,7 @@ <span class="sourceLineNo">019</span><a name="line.19"></a> <span class="sourceLineNo">020</span>import org.apache.hadoop.hbase.classification.InterfaceAudience;<a name="line.20"></a> <span class="sourceLineNo">021</span>import org.apache.hadoop.hbase.classification.InterfaceStability;<a name="line.21"></a> -<span class="sourceLineNo">022</span>import org.apache.hadoop.hbase.protobuf.generated.HBaseProtos.SnapshotDescription;<a name="line.22"></a> +<span class="sourceLineNo">022</span>import org.apache.hadoop.hbase.client.SnapshotDescription;<a name="line.22"></a> <span class="sourceLineNo">023</span><a name="line.23"></a> <span class="sourceLineNo">024</span>/**<a name="line.24"></a> <span class="sourceLineNo">025</span> * Thrown when a snapshot exists but should not<a name="line.25"></a>