| <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd"> |
| <html lang="en"> |
| <head> |
| <title>Source code</title> |
| <link rel="stylesheet" type="text/css" href="../../../../../../../stylesheet.css" title="Style"> |
| </head> |
| <body> |
| <div class="sourceContainer"> |
| <pre><span class="sourceLineNo">001</span>/**<a name="line.1"></a> |
| <span class="sourceLineNo">002</span> * Licensed to the Apache Software Foundation (ASF) under one<a name="line.2"></a> |
| <span class="sourceLineNo">003</span> * or more contributor license agreements. See the NOTICE file<a name="line.3"></a> |
| <span class="sourceLineNo">004</span> * distributed with this work for additional information<a name="line.4"></a> |
| <span class="sourceLineNo">005</span> * regarding copyright ownership. The ASF licenses this file<a name="line.5"></a> |
| <span class="sourceLineNo">006</span> * to you under the Apache License, Version 2.0 (the<a name="line.6"></a> |
| <span class="sourceLineNo">007</span> * "License"); you may not use this file except in compliance<a name="line.7"></a> |
| <span class="sourceLineNo">008</span> * with the License. You may obtain a copy of the License at<a name="line.8"></a> |
| <span class="sourceLineNo">009</span> *<a name="line.9"></a> |
| <span class="sourceLineNo">010</span> * http://www.apache.org/licenses/LICENSE-2.0<a name="line.10"></a> |
| <span class="sourceLineNo">011</span> *<a name="line.11"></a> |
| <span class="sourceLineNo">012</span> * Unless required by applicable law or agreed to in writing, software<a name="line.12"></a> |
| <span class="sourceLineNo">013</span> * distributed under the License is distributed on an "AS IS" BASIS,<a name="line.13"></a> |
| <span class="sourceLineNo">014</span> * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.<a name="line.14"></a> |
| <span class="sourceLineNo">015</span> * See the License for the specific language governing permissions and<a name="line.15"></a> |
| <span class="sourceLineNo">016</span> * limitations under the License.<a name="line.16"></a> |
| <span class="sourceLineNo">017</span> */<a name="line.17"></a> |
| <span class="sourceLineNo">018</span>package org.apache.hadoop.hbase.master.snapshot;<a name="line.18"></a> |
| <span class="sourceLineNo">019</span><a name="line.19"></a> |
| <span class="sourceLineNo">020</span>import java.io.IOException;<a name="line.20"></a> |
| <span class="sourceLineNo">021</span>import java.util.Collection;<a name="line.21"></a> |
| <span class="sourceLineNo">022</span>import java.util.HashMap;<a name="line.22"></a> |
| <span class="sourceLineNo">023</span>import java.util.HashSet;<a name="line.23"></a> |
| <span class="sourceLineNo">024</span>import java.util.List;<a name="line.24"></a> |
| <span class="sourceLineNo">025</span>import java.util.Map;<a name="line.25"></a> |
| <span class="sourceLineNo">026</span>import java.util.Set;<a name="line.26"></a> |
| <span class="sourceLineNo">027</span>import java.util.Timer;<a name="line.27"></a> |
| <span class="sourceLineNo">028</span>import java.util.TimerTask;<a name="line.28"></a> |
| <span class="sourceLineNo">029</span>import java.util.concurrent.locks.Lock;<a name="line.29"></a> |
| <span class="sourceLineNo">030</span>import org.apache.commons.lang3.ArrayUtils;<a name="line.30"></a> |
| <span class="sourceLineNo">031</span>import org.apache.hadoop.conf.Configuration;<a name="line.31"></a> |
| <span class="sourceLineNo">032</span>import org.apache.hadoop.fs.FileStatus;<a name="line.32"></a> |
| <span class="sourceLineNo">033</span>import org.apache.hadoop.fs.FileSystem;<a name="line.33"></a> |
| <span class="sourceLineNo">034</span>import org.apache.hadoop.fs.Path;<a name="line.34"></a> |
| <span class="sourceLineNo">035</span>import org.apache.hadoop.hbase.Stoppable;<a name="line.35"></a> |
| <span class="sourceLineNo">036</span>import org.apache.hadoop.hbase.snapshot.CorruptedSnapshotException;<a name="line.36"></a> |
| <span class="sourceLineNo">037</span>import org.apache.hadoop.hbase.snapshot.SnapshotDescriptionUtils;<a name="line.37"></a> |
| <span class="sourceLineNo">038</span>import org.apache.hadoop.hbase.util.CommonFSUtils;<a name="line.38"></a> |
| <span class="sourceLineNo">039</span>import org.apache.yetus.audience.InterfaceAudience;<a name="line.39"></a> |
| <span class="sourceLineNo">040</span>import org.apache.yetus.audience.InterfaceStability;<a name="line.40"></a> |
| <span class="sourceLineNo">041</span>import org.slf4j.Logger;<a name="line.41"></a> |
| <span class="sourceLineNo">042</span>import org.slf4j.LoggerFactory;<a name="line.42"></a> |
| <span class="sourceLineNo">043</span><a name="line.43"></a> |
| <span class="sourceLineNo">044</span>import org.apache.hbase.thirdparty.com.google.common.annotations.VisibleForTesting;<a name="line.44"></a> |
| <span class="sourceLineNo">045</span>import org.apache.hbase.thirdparty.com.google.common.collect.Lists;<a name="line.45"></a> |
| <span class="sourceLineNo">046</span><a name="line.46"></a> |
| <span class="sourceLineNo">047</span>/**<a name="line.47"></a> |
| <span class="sourceLineNo">048</span> * Intelligently keep track of all the files for all the snapshots.<a name="line.48"></a> |
| <span class="sourceLineNo">049</span> * <p><a name="line.49"></a> |
| <span class="sourceLineNo">050</span> * A cache of files is kept to avoid querying the {@link FileSystem} frequently. If there is a cache<a name="line.50"></a> |
| <span class="sourceLineNo">051</span> * miss the directory modification time is used to ensure that we don't rescan directories that we<a name="line.51"></a> |
| <span class="sourceLineNo">052</span> * already have in cache. We only check the modification times of the snapshot directories<a name="line.52"></a> |
| <span class="sourceLineNo">053</span> * (/hbase/.snapshot/[snapshot_name]) to determine if the files need to be loaded into the cache.<a name="line.53"></a> |
| <span class="sourceLineNo">054</span> * <p><a name="line.54"></a> |
| <span class="sourceLineNo">055</span> * New snapshots will be added to the cache and deleted snapshots will be removed when we refresh<a name="line.55"></a> |
| <span class="sourceLineNo">056</span> * the cache. If the files underneath a snapshot directory are changed, but not the snapshot itself,<a name="line.56"></a> |
| <span class="sourceLineNo">057</span> * we will ignore updates to that snapshot's files.<a name="line.57"></a> |
| <span class="sourceLineNo">058</span> * <p><a name="line.58"></a> |
| <span class="sourceLineNo">059</span> * This is sufficient because each snapshot has its own directory and is added via an atomic rename<a name="line.59"></a> |
| <span class="sourceLineNo">060</span> * <i>once</i>, when the snapshot is created. We don't need to worry about the data in the snapshot<a name="line.60"></a> |
| <span class="sourceLineNo">061</span> * being run.<a name="line.61"></a> |
| <span class="sourceLineNo">062</span> * <p><a name="line.62"></a> |
| <span class="sourceLineNo">063</span> * Further, the cache is periodically refreshed ensure that files in snapshots that were deleted are<a name="line.63"></a> |
| <span class="sourceLineNo">064</span> * also removed from the cache.<a name="line.64"></a> |
| <span class="sourceLineNo">065</span> * <p><a name="line.65"></a> |
| <span class="sourceLineNo">066</span> * A {@link SnapshotFileCache.SnapshotFileInspector} must be passed when creating <tt>this</tt> to<a name="line.66"></a> |
| <span class="sourceLineNo">067</span> * allow extraction of files under /hbase/.snapshot/[snapshot name] directory, for each snapshot.<a name="line.67"></a> |
| <span class="sourceLineNo">068</span> * This allows you to only cache files under, for instance, all the logs in the .logs directory or<a name="line.68"></a> |
| <span class="sourceLineNo">069</span> * all the files under all the regions.<a name="line.69"></a> |
| <span class="sourceLineNo">070</span> * <p><a name="line.70"></a> |
| <span class="sourceLineNo">071</span> * <tt>this</tt> also considers all running snapshots (those under /hbase/.snapshot/.tmp) as valid<a name="line.71"></a> |
| <span class="sourceLineNo">072</span> * snapshots and will attempt to cache files from those snapshots as well.<a name="line.72"></a> |
| <span class="sourceLineNo">073</span> * <p><a name="line.73"></a> |
| <span class="sourceLineNo">074</span> * Queries about a given file are thread-safe with respect to multiple queries and cache refreshes.<a name="line.74"></a> |
| <span class="sourceLineNo">075</span> */<a name="line.75"></a> |
| <span class="sourceLineNo">076</span>@InterfaceAudience.Private<a name="line.76"></a> |
| <span class="sourceLineNo">077</span>@InterfaceStability.Evolving<a name="line.77"></a> |
| <span class="sourceLineNo">078</span>public class SnapshotFileCache implements Stoppable {<a name="line.78"></a> |
| <span class="sourceLineNo">079</span> interface SnapshotFileInspector {<a name="line.79"></a> |
| <span class="sourceLineNo">080</span> /**<a name="line.80"></a> |
| <span class="sourceLineNo">081</span> * Returns a collection of file names needed by the snapshot.<a name="line.81"></a> |
| <span class="sourceLineNo">082</span> * @param fs {@link FileSystem} where snapshot mainifest files are stored<a name="line.82"></a> |
| <span class="sourceLineNo">083</span> * @param snapshotDir {@link Path} to the snapshot directory to scan.<a name="line.83"></a> |
| <span class="sourceLineNo">084</span> * @return the collection of file names needed by the snapshot.<a name="line.84"></a> |
| <span class="sourceLineNo">085</span> */<a name="line.85"></a> |
| <span class="sourceLineNo">086</span> Collection<String> filesUnderSnapshot(final FileSystem fs, final Path snapshotDir)<a name="line.86"></a> |
| <span class="sourceLineNo">087</span> throws IOException;<a name="line.87"></a> |
| <span class="sourceLineNo">088</span> }<a name="line.88"></a> |
| <span class="sourceLineNo">089</span><a name="line.89"></a> |
| <span class="sourceLineNo">090</span> private static final Logger LOG = LoggerFactory.getLogger(SnapshotFileCache.class);<a name="line.90"></a> |
| <span class="sourceLineNo">091</span> private volatile boolean stop = false;<a name="line.91"></a> |
| <span class="sourceLineNo">092</span> private final FileSystem fs, workingFs;<a name="line.92"></a> |
| <span class="sourceLineNo">093</span> private final SnapshotFileInspector fileInspector;<a name="line.93"></a> |
| <span class="sourceLineNo">094</span> private final Path snapshotDir, workingSnapshotDir;<a name="line.94"></a> |
| <span class="sourceLineNo">095</span> private final Set<String> cache = new HashSet<>();<a name="line.95"></a> |
| <span class="sourceLineNo">096</span> /**<a name="line.96"></a> |
| <span class="sourceLineNo">097</span> * This is a helper map of information about the snapshot directories so we don't need to rescan<a name="line.97"></a> |
| <span class="sourceLineNo">098</span> * them if they haven't changed since the last time we looked.<a name="line.98"></a> |
| <span class="sourceLineNo">099</span> */<a name="line.99"></a> |
| <span class="sourceLineNo">100</span> private final Map<String, SnapshotDirectoryInfo> snapshots = new HashMap<>();<a name="line.100"></a> |
| <span class="sourceLineNo">101</span> private final Timer refreshTimer;<a name="line.101"></a> |
| <span class="sourceLineNo">102</span><a name="line.102"></a> |
| <span class="sourceLineNo">103</span> /**<a name="line.103"></a> |
| <span class="sourceLineNo">104</span> * Create a snapshot file cache for all snapshots under the specified [root]/.snapshot on the<a name="line.104"></a> |
| <span class="sourceLineNo">105</span> * filesystem.<a name="line.105"></a> |
| <span class="sourceLineNo">106</span> * <p><a name="line.106"></a> |
| <span class="sourceLineNo">107</span> * Immediately loads the file cache.<a name="line.107"></a> |
| <span class="sourceLineNo">108</span> * @param conf to extract the configured {@link FileSystem} where the snapshots are stored and<a name="line.108"></a> |
| <span class="sourceLineNo">109</span> * hbase root directory<a name="line.109"></a> |
| <span class="sourceLineNo">110</span> * @param cacheRefreshPeriod frequency (ms) with which the cache should be refreshed<a name="line.110"></a> |
| <span class="sourceLineNo">111</span> * @param cacheRefreshDelay amount of time to wait for the cache to be refreshed<a name="line.111"></a> |
| <span class="sourceLineNo">112</span> * @param refreshThreadName name of the cache refresh thread<a name="line.112"></a> |
| <span class="sourceLineNo">113</span> * @param inspectSnapshotFiles Filter to apply to each snapshot to extract the files.<a name="line.113"></a> |
| <span class="sourceLineNo">114</span> * @throws IOException if the {@link FileSystem} or root directory cannot be loaded<a name="line.114"></a> |
| <span class="sourceLineNo">115</span> */<a name="line.115"></a> |
| <span class="sourceLineNo">116</span> public SnapshotFileCache(Configuration conf, long cacheRefreshPeriod, long cacheRefreshDelay,<a name="line.116"></a> |
| <span class="sourceLineNo">117</span> String refreshThreadName, SnapshotFileInspector inspectSnapshotFiles) throws IOException {<a name="line.117"></a> |
| <span class="sourceLineNo">118</span> this(CommonFSUtils.getCurrentFileSystem(conf), CommonFSUtils.getRootDir(conf),<a name="line.118"></a> |
| <span class="sourceLineNo">119</span> SnapshotDescriptionUtils.getWorkingSnapshotDir(CommonFSUtils.getRootDir(conf), conf).<a name="line.119"></a> |
| <span class="sourceLineNo">120</span> getFileSystem(conf),<a name="line.120"></a> |
| <span class="sourceLineNo">121</span> SnapshotDescriptionUtils.getWorkingSnapshotDir(CommonFSUtils.getRootDir(conf), conf),<a name="line.121"></a> |
| <span class="sourceLineNo">122</span> cacheRefreshPeriod, cacheRefreshDelay, refreshThreadName, inspectSnapshotFiles);<a name="line.122"></a> |
| <span class="sourceLineNo">123</span> }<a name="line.123"></a> |
| <span class="sourceLineNo">124</span><a name="line.124"></a> |
| <span class="sourceLineNo">125</span> /**<a name="line.125"></a> |
| <span class="sourceLineNo">126</span> * Create a snapshot file cache for all snapshots under the specified [root]/.snapshot on the<a name="line.126"></a> |
| <span class="sourceLineNo">127</span> * filesystem<a name="line.127"></a> |
| <span class="sourceLineNo">128</span> * @param fs {@link FileSystem} where the snapshots are stored<a name="line.128"></a> |
| <span class="sourceLineNo">129</span> * @param rootDir hbase root directory<a name="line.129"></a> |
| <span class="sourceLineNo">130</span> * @param workingFs {@link FileSystem} where ongoing snapshot mainifest files are stored<a name="line.130"></a> |
| <span class="sourceLineNo">131</span> * @param workingDir Location to store ongoing snapshot manifest files<a name="line.131"></a> |
| <span class="sourceLineNo">132</span> * @param cacheRefreshPeriod period (ms) with which the cache should be refreshed<a name="line.132"></a> |
| <span class="sourceLineNo">133</span> * @param cacheRefreshDelay amount of time to wait for the cache to be refreshed<a name="line.133"></a> |
| <span class="sourceLineNo">134</span> * @param refreshThreadName name of the cache refresh thread<a name="line.134"></a> |
| <span class="sourceLineNo">135</span> * @param inspectSnapshotFiles Filter to apply to each snapshot to extract the files.<a name="line.135"></a> |
| <span class="sourceLineNo">136</span> */<a name="line.136"></a> |
| <span class="sourceLineNo">137</span> public SnapshotFileCache(FileSystem fs, Path rootDir, FileSystem workingFs, Path workingDir,<a name="line.137"></a> |
| <span class="sourceLineNo">138</span> long cacheRefreshPeriod, long cacheRefreshDelay, String refreshThreadName,<a name="line.138"></a> |
| <span class="sourceLineNo">139</span> SnapshotFileInspector inspectSnapshotFiles) {<a name="line.139"></a> |
| <span class="sourceLineNo">140</span> this.fs = fs;<a name="line.140"></a> |
| <span class="sourceLineNo">141</span> this.workingFs = workingFs;<a name="line.141"></a> |
| <span class="sourceLineNo">142</span> this.workingSnapshotDir = workingDir;<a name="line.142"></a> |
| <span class="sourceLineNo">143</span> this.fileInspector = inspectSnapshotFiles;<a name="line.143"></a> |
| <span class="sourceLineNo">144</span> this.snapshotDir = SnapshotDescriptionUtils.getSnapshotsDir(rootDir);<a name="line.144"></a> |
| <span class="sourceLineNo">145</span> // periodically refresh the file cache to make sure we aren't superfluously saving files.<a name="line.145"></a> |
| <span class="sourceLineNo">146</span> this.refreshTimer = new Timer(refreshThreadName, true);<a name="line.146"></a> |
| <span class="sourceLineNo">147</span> this.refreshTimer.scheduleAtFixedRate(new RefreshCacheTask(), cacheRefreshDelay,<a name="line.147"></a> |
| <span class="sourceLineNo">148</span> cacheRefreshPeriod);<a name="line.148"></a> |
| <span class="sourceLineNo">149</span> }<a name="line.149"></a> |
| <span class="sourceLineNo">150</span><a name="line.150"></a> |
| <span class="sourceLineNo">151</span> /**<a name="line.151"></a> |
| <span class="sourceLineNo">152</span> * Trigger a cache refresh, even if its before the next cache refresh. Does not affect pending<a name="line.152"></a> |
| <span class="sourceLineNo">153</span> * cache refreshes.<a name="line.153"></a> |
| <span class="sourceLineNo">154</span> * <p/><a name="line.154"></a> |
| <span class="sourceLineNo">155</span> * Blocks until the cache is refreshed.<a name="line.155"></a> |
| <span class="sourceLineNo">156</span> * <p/><a name="line.156"></a> |
| <span class="sourceLineNo">157</span> * Exposed for TESTING.<a name="line.157"></a> |
| <span class="sourceLineNo">158</span> */<a name="line.158"></a> |
| <span class="sourceLineNo">159</span> public synchronized void triggerCacheRefreshForTesting() {<a name="line.159"></a> |
| <span class="sourceLineNo">160</span> try {<a name="line.160"></a> |
| <span class="sourceLineNo">161</span> refreshCache();<a name="line.161"></a> |
| <span class="sourceLineNo">162</span> } catch (IOException e) {<a name="line.162"></a> |
| <span class="sourceLineNo">163</span> LOG.warn("Failed to refresh snapshot hfile cache!", e);<a name="line.163"></a> |
| <span class="sourceLineNo">164</span> }<a name="line.164"></a> |
| <span class="sourceLineNo">165</span> LOG.debug("Current cache:" + cache);<a name="line.165"></a> |
| <span class="sourceLineNo">166</span> }<a name="line.166"></a> |
| <span class="sourceLineNo">167</span><a name="line.167"></a> |
| <span class="sourceLineNo">168</span> /**<a name="line.168"></a> |
| <span class="sourceLineNo">169</span> * Check to see if any of the passed file names is contained in any of the snapshots. First checks<a name="line.169"></a> |
| <span class="sourceLineNo">170</span> * an in-memory cache of the files to keep. If its not in the cache, then the cache is refreshed<a name="line.170"></a> |
| <span class="sourceLineNo">171</span> * and the cache checked again for that file. This ensures that we never return files that exist.<a name="line.171"></a> |
| <span class="sourceLineNo">172</span> * <p><a name="line.172"></a> |
| <span class="sourceLineNo">173</span> * Note this may lead to periodic false positives for the file being referenced. Periodically, the<a name="line.173"></a> |
| <span class="sourceLineNo">174</span> * cache is refreshed even if there are no requests to ensure that the false negatives get removed<a name="line.174"></a> |
| <span class="sourceLineNo">175</span> * eventually. For instance, suppose you have a file in the snapshot and it gets loaded into the<a name="line.175"></a> |
| <span class="sourceLineNo">176</span> * cache. Then at some point later that snapshot is deleted. If the cache has not been refreshed<a name="line.176"></a> |
| <span class="sourceLineNo">177</span> * at that point, cache will still think the file system contains that file and return<a name="line.177"></a> |
| <span class="sourceLineNo">178</span> * <tt>true</tt>, even if it is no longer present (false positive). However, if the file never was<a name="line.178"></a> |
| <span class="sourceLineNo">179</span> * on the filesystem, we will never find it and always return <tt>false</tt>.<a name="line.179"></a> |
| <span class="sourceLineNo">180</span> * @param files file to check, NOTE: Relies that files are loaded from hdfs before method is<a name="line.180"></a> |
| <span class="sourceLineNo">181</span> * called (NOT LAZY)<a name="line.181"></a> |
| <span class="sourceLineNo">182</span> * @return <tt>unReferencedFiles</tt> the collection of files that do not have snapshot references<a name="line.182"></a> |
| <span class="sourceLineNo">183</span> * @throws IOException if there is an unexpected error reaching the filesystem.<a name="line.183"></a> |
| <span class="sourceLineNo">184</span> */<a name="line.184"></a> |
| <span class="sourceLineNo">185</span> // XXX this is inefficient to synchronize on the method, when what we really need to guard against<a name="line.185"></a> |
| <span class="sourceLineNo">186</span> // is an illegal access to the cache. Really we could do a mutex-guarded pointer swap on the<a name="line.186"></a> |
| <span class="sourceLineNo">187</span> // cache, but that seems overkill at the moment and isn't necessarily a bottleneck.<a name="line.187"></a> |
| <span class="sourceLineNo">188</span> public synchronized Iterable<FileStatus> getUnreferencedFiles(Iterable<FileStatus> files,<a name="line.188"></a> |
| <span class="sourceLineNo">189</span> final SnapshotManager snapshotManager) throws IOException {<a name="line.189"></a> |
| <span class="sourceLineNo">190</span> List<FileStatus> unReferencedFiles = Lists.newArrayList();<a name="line.190"></a> |
| <span class="sourceLineNo">191</span> List<String> snapshotsInProgress = null;<a name="line.191"></a> |
| <span class="sourceLineNo">192</span> boolean refreshed = false;<a name="line.192"></a> |
| <span class="sourceLineNo">193</span> Lock lock = null;<a name="line.193"></a> |
| <span class="sourceLineNo">194</span> if (snapshotManager != null) {<a name="line.194"></a> |
| <span class="sourceLineNo">195</span> lock = snapshotManager.getTakingSnapshotLock().writeLock();<a name="line.195"></a> |
| <span class="sourceLineNo">196</span> }<a name="line.196"></a> |
| <span class="sourceLineNo">197</span> if (lock == null || lock.tryLock()) {<a name="line.197"></a> |
| <span class="sourceLineNo">198</span> try {<a name="line.198"></a> |
| <span class="sourceLineNo">199</span> if (snapshotManager != null && snapshotManager.isTakingAnySnapshot()) {<a name="line.199"></a> |
| <span class="sourceLineNo">200</span> LOG.warn("Not checking unreferenced files since snapshot is running, it will " +<a name="line.200"></a> |
| <span class="sourceLineNo">201</span> "skip to clean the HFiles this time");<a name="line.201"></a> |
| <span class="sourceLineNo">202</span> return unReferencedFiles;<a name="line.202"></a> |
| <span class="sourceLineNo">203</span> }<a name="line.203"></a> |
| <span class="sourceLineNo">204</span> for (FileStatus file : files) {<a name="line.204"></a> |
| <span class="sourceLineNo">205</span> String fileName = file.getPath().getName();<a name="line.205"></a> |
| <span class="sourceLineNo">206</span> if (!refreshed && !cache.contains(fileName)) {<a name="line.206"></a> |
| <span class="sourceLineNo">207</span> refreshCache();<a name="line.207"></a> |
| <span class="sourceLineNo">208</span> refreshed = true;<a name="line.208"></a> |
| <span class="sourceLineNo">209</span> }<a name="line.209"></a> |
| <span class="sourceLineNo">210</span> if (cache.contains(fileName)) {<a name="line.210"></a> |
| <span class="sourceLineNo">211</span> continue;<a name="line.211"></a> |
| <span class="sourceLineNo">212</span> }<a name="line.212"></a> |
| <span class="sourceLineNo">213</span> if (snapshotsInProgress == null) {<a name="line.213"></a> |
| <span class="sourceLineNo">214</span> snapshotsInProgress = getSnapshotsInProgress();<a name="line.214"></a> |
| <span class="sourceLineNo">215</span> }<a name="line.215"></a> |
| <span class="sourceLineNo">216</span> if (snapshotsInProgress.contains(fileName)) {<a name="line.216"></a> |
| <span class="sourceLineNo">217</span> continue;<a name="line.217"></a> |
| <span class="sourceLineNo">218</span> }<a name="line.218"></a> |
| <span class="sourceLineNo">219</span> unReferencedFiles.add(file);<a name="line.219"></a> |
| <span class="sourceLineNo">220</span> }<a name="line.220"></a> |
| <span class="sourceLineNo">221</span> } finally {<a name="line.221"></a> |
| <span class="sourceLineNo">222</span> if (lock != null) {<a name="line.222"></a> |
| <span class="sourceLineNo">223</span> lock.unlock();<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> return unReferencedFiles;<a name="line.227"></a> |
| <span class="sourceLineNo">228</span> }<a name="line.228"></a> |
| <span class="sourceLineNo">229</span><a name="line.229"></a> |
| <span class="sourceLineNo">230</span> private void refreshCache() throws IOException {<a name="line.230"></a> |
| <span class="sourceLineNo">231</span> // just list the snapshot directory directly, do not check the modification time for the root<a name="line.231"></a> |
| <span class="sourceLineNo">232</span> // snapshot directory, as some file system implementations do not modify the parent directory's<a name="line.232"></a> |
| <span class="sourceLineNo">233</span> // modTime when there are new sub items, for example, S3.<a name="line.233"></a> |
| <span class="sourceLineNo">234</span> FileStatus[] snapshotDirs = CommonFSUtils.listStatus(fs, snapshotDir,<a name="line.234"></a> |
| <span class="sourceLineNo">235</span> p -> !p.getName().equals(SnapshotDescriptionUtils.SNAPSHOT_TMP_DIR_NAME));<a name="line.235"></a> |
| <span class="sourceLineNo">236</span><a name="line.236"></a> |
| <span class="sourceLineNo">237</span> // clear the cache, as in the below code, either we will also clear the snapshots, or we will<a name="line.237"></a> |
| <span class="sourceLineNo">238</span> // refill the file name cache again.<a name="line.238"></a> |
| <span class="sourceLineNo">239</span> this.cache.clear();<a name="line.239"></a> |
| <span class="sourceLineNo">240</span> if (ArrayUtils.isEmpty(snapshotDirs)) {<a name="line.240"></a> |
| <span class="sourceLineNo">241</span> // remove all the remembered snapshots because we don't have any left<a name="line.241"></a> |
| <span class="sourceLineNo">242</span> if (LOG.isDebugEnabled() && this.snapshots.size() > 0) {<a name="line.242"></a> |
| <span class="sourceLineNo">243</span> LOG.debug("No snapshots on-disk, clear cache");<a name="line.243"></a> |
| <span class="sourceLineNo">244</span> }<a name="line.244"></a> |
| <span class="sourceLineNo">245</span> this.snapshots.clear();<a name="line.245"></a> |
| <span class="sourceLineNo">246</span> return;<a name="line.246"></a> |
| <span class="sourceLineNo">247</span> }<a name="line.247"></a> |
| <span class="sourceLineNo">248</span><a name="line.248"></a> |
| <span class="sourceLineNo">249</span> // iterate over all the cached snapshots and see if we need to update some, it is not an<a name="line.249"></a> |
| <span class="sourceLineNo">250</span> // expensive operation if we do not reload the manifest of snapshots.<a name="line.250"></a> |
| <span class="sourceLineNo">251</span> Map<String, SnapshotDirectoryInfo> newSnapshots = new HashMap<>();<a name="line.251"></a> |
| <span class="sourceLineNo">252</span> for (FileStatus snapshotDir : snapshotDirs) {<a name="line.252"></a> |
| <span class="sourceLineNo">253</span> String name = snapshotDir.getPath().getName();<a name="line.253"></a> |
| <span class="sourceLineNo">254</span> SnapshotDirectoryInfo files = this.snapshots.remove(name);<a name="line.254"></a> |
| <span class="sourceLineNo">255</span> // if we don't know about the snapshot or its been modified, we need to update the<a name="line.255"></a> |
| <span class="sourceLineNo">256</span> // files the latter could occur where I create a snapshot, then delete it, and then make a<a name="line.256"></a> |
| <span class="sourceLineNo">257</span> // new snapshot with the same name. We will need to update the cache the information from<a name="line.257"></a> |
| <span class="sourceLineNo">258</span> // that new snapshot, even though it has the same name as the files referenced have<a name="line.258"></a> |
| <span class="sourceLineNo">259</span> // probably changed.<a name="line.259"></a> |
| <span class="sourceLineNo">260</span> if (files == null || files.hasBeenModified(snapshotDir.getModificationTime())) {<a name="line.260"></a> |
| <span class="sourceLineNo">261</span> Collection<String> storedFiles = fileInspector.filesUnderSnapshot(fs,<a name="line.261"></a> |
| <span class="sourceLineNo">262</span> snapshotDir.getPath());<a name="line.262"></a> |
| <span class="sourceLineNo">263</span> files = new SnapshotDirectoryInfo(snapshotDir.getModificationTime(), storedFiles);<a name="line.263"></a> |
| <span class="sourceLineNo">264</span> }<a name="line.264"></a> |
| <span class="sourceLineNo">265</span> // add all the files to cache<a name="line.265"></a> |
| <span class="sourceLineNo">266</span> this.cache.addAll(files.getFiles());<a name="line.266"></a> |
| <span class="sourceLineNo">267</span> newSnapshots.put(name, files);<a name="line.267"></a> |
| <span class="sourceLineNo">268</span> }<a name="line.268"></a> |
| <span class="sourceLineNo">269</span> // set the snapshots we are tracking<a name="line.269"></a> |
| <span class="sourceLineNo">270</span> this.snapshots.clear();<a name="line.270"></a> |
| <span class="sourceLineNo">271</span> this.snapshots.putAll(newSnapshots);<a name="line.271"></a> |
| <span class="sourceLineNo">272</span> }<a name="line.272"></a> |
| <span class="sourceLineNo">273</span><a name="line.273"></a> |
| <span class="sourceLineNo">274</span> @VisibleForTesting<a name="line.274"></a> |
| <span class="sourceLineNo">275</span> List<String> getSnapshotsInProgress() throws IOException {<a name="line.275"></a> |
| <span class="sourceLineNo">276</span> List<String> snapshotInProgress = Lists.newArrayList();<a name="line.276"></a> |
| <span class="sourceLineNo">277</span> // only add those files to the cache, but not to the known snapshots<a name="line.277"></a> |
| <span class="sourceLineNo">278</span><a name="line.278"></a> |
| <span class="sourceLineNo">279</span> FileStatus[] snapshotsInProgress = CommonFSUtils.listStatus(this.workingFs, this.workingSnapshotDir);<a name="line.279"></a> |
| <span class="sourceLineNo">280</span><a name="line.280"></a> |
| <span class="sourceLineNo">281</span> if (!ArrayUtils.isEmpty(snapshotsInProgress)) {<a name="line.281"></a> |
| <span class="sourceLineNo">282</span> for (FileStatus snapshot : snapshotsInProgress) {<a name="line.282"></a> |
| <span class="sourceLineNo">283</span> try {<a name="line.283"></a> |
| <span class="sourceLineNo">284</span> snapshotInProgress.addAll(fileInspector.filesUnderSnapshot(workingFs,<a name="line.284"></a> |
| <span class="sourceLineNo">285</span> snapshot.getPath()));<a name="line.285"></a> |
| <span class="sourceLineNo">286</span> } catch (CorruptedSnapshotException cse) {<a name="line.286"></a> |
| <span class="sourceLineNo">287</span> LOG.info("Corrupted in-progress snapshot file exception, ignored.", cse);<a name="line.287"></a> |
| <span class="sourceLineNo">288</span> }<a name="line.288"></a> |
| <span class="sourceLineNo">289</span> }<a name="line.289"></a> |
| <span class="sourceLineNo">290</span> }<a name="line.290"></a> |
| <span class="sourceLineNo">291</span> return snapshotInProgress;<a name="line.291"></a> |
| <span class="sourceLineNo">292</span> }<a name="line.292"></a> |
| <span class="sourceLineNo">293</span><a name="line.293"></a> |
| <span class="sourceLineNo">294</span> /**<a name="line.294"></a> |
| <span class="sourceLineNo">295</span> * Simple helper task that just periodically attempts to refresh the cache<a name="line.295"></a> |
| <span class="sourceLineNo">296</span> */<a name="line.296"></a> |
| <span class="sourceLineNo">297</span> public class RefreshCacheTask extends TimerTask {<a name="line.297"></a> |
| <span class="sourceLineNo">298</span> @Override<a name="line.298"></a> |
| <span class="sourceLineNo">299</span> public void run() {<a name="line.299"></a> |
| <span class="sourceLineNo">300</span> synchronized (SnapshotFileCache.this) {<a name="line.300"></a> |
| <span class="sourceLineNo">301</span> try {<a name="line.301"></a> |
| <span class="sourceLineNo">302</span> SnapshotFileCache.this.refreshCache();<a name="line.302"></a> |
| <span class="sourceLineNo">303</span> } catch (IOException e) {<a name="line.303"></a> |
| <span class="sourceLineNo">304</span> LOG.warn("Failed to refresh snapshot hfile cache!", e);<a name="line.304"></a> |
| <span class="sourceLineNo">305</span> // clear all the cached entries if we meet an error<a name="line.305"></a> |
| <span class="sourceLineNo">306</span> cache.clear();<a name="line.306"></a> |
| <span class="sourceLineNo">307</span> snapshots.clear();<a name="line.307"></a> |
| <span class="sourceLineNo">308</span> }<a name="line.308"></a> |
| <span class="sourceLineNo">309</span> }<a name="line.309"></a> |
| <span class="sourceLineNo">310</span><a name="line.310"></a> |
| <span class="sourceLineNo">311</span> }<a name="line.311"></a> |
| <span class="sourceLineNo">312</span> }<a name="line.312"></a> |
| <span class="sourceLineNo">313</span><a name="line.313"></a> |
| <span class="sourceLineNo">314</span> @Override<a name="line.314"></a> |
| <span class="sourceLineNo">315</span> public void stop(String why) {<a name="line.315"></a> |
| <span class="sourceLineNo">316</span> if (!this.stop) {<a name="line.316"></a> |
| <span class="sourceLineNo">317</span> this.stop = true;<a name="line.317"></a> |
| <span class="sourceLineNo">318</span> this.refreshTimer.cancel();<a name="line.318"></a> |
| <span class="sourceLineNo">319</span> }<a name="line.319"></a> |
| <span class="sourceLineNo">320</span><a name="line.320"></a> |
| <span class="sourceLineNo">321</span> }<a name="line.321"></a> |
| <span class="sourceLineNo">322</span><a name="line.322"></a> |
| <span class="sourceLineNo">323</span> @Override<a name="line.323"></a> |
| <span class="sourceLineNo">324</span> public boolean isStopped() {<a name="line.324"></a> |
| <span class="sourceLineNo">325</span> return this.stop;<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> /**<a name="line.328"></a> |
| <span class="sourceLineNo">329</span> * Information about a snapshot directory<a name="line.329"></a> |
| <span class="sourceLineNo">330</span> */<a name="line.330"></a> |
| <span class="sourceLineNo">331</span> private static class SnapshotDirectoryInfo {<a name="line.331"></a> |
| <span class="sourceLineNo">332</span> long lastModified;<a name="line.332"></a> |
| <span class="sourceLineNo">333</span> Collection<String> files;<a name="line.333"></a> |
| <span class="sourceLineNo">334</span><a name="line.334"></a> |
| <span class="sourceLineNo">335</span> public SnapshotDirectoryInfo(long mtime, Collection<String> files) {<a name="line.335"></a> |
| <span class="sourceLineNo">336</span> this.lastModified = mtime;<a name="line.336"></a> |
| <span class="sourceLineNo">337</span> this.files = files;<a name="line.337"></a> |
| <span class="sourceLineNo">338</span> }<a name="line.338"></a> |
| <span class="sourceLineNo">339</span><a name="line.339"></a> |
| <span class="sourceLineNo">340</span> /**<a name="line.340"></a> |
| <span class="sourceLineNo">341</span> * @return the hfiles in the snapshot when <tt>this</tt> was made.<a name="line.341"></a> |
| <span class="sourceLineNo">342</span> */<a name="line.342"></a> |
| <span class="sourceLineNo">343</span> public Collection<String> getFiles() {<a name="line.343"></a> |
| <span class="sourceLineNo">344</span> return this.files;<a name="line.344"></a> |
| <span class="sourceLineNo">345</span> }<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">348</span> * Check if the snapshot directory has been modified<a name="line.348"></a> |
| <span class="sourceLineNo">349</span> * @param mtime current modification time of the directory<a name="line.349"></a> |
| <span class="sourceLineNo">350</span> * @return <tt>true</tt> if it the modification time of the directory is newer time when we<a name="line.350"></a> |
| <span class="sourceLineNo">351</span> * created <tt>this</tt><a name="line.351"></a> |
| <span class="sourceLineNo">352</span> */<a name="line.352"></a> |
| <span class="sourceLineNo">353</span> public boolean hasBeenModified(long mtime) {<a name="line.353"></a> |
| <span class="sourceLineNo">354</span> return this.lastModified < mtime;<a name="line.354"></a> |
| <span class="sourceLineNo">355</span> }<a name="line.355"></a> |
| <span class="sourceLineNo">356</span> }<a name="line.356"></a> |
| <span class="sourceLineNo">357</span>}<a name="line.357"></a> |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| </pre> |
| </div> |
| </body> |
| </html> |
