1<?xml version="1.0" encoding="UTF-8" standalone="no"?> 2<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> 3<html xmlns="http://www.w3.org/1999/xhtml"> 4 <head> 5 <meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /> 6 <title>Backup Procedures</title> 7 <link rel="stylesheet" href="gettingStarted.css" type="text/css" /> 8 <meta name="generator" content="DocBook XSL Stylesheets V1.73.2" /> 9 <link rel="start" href="index.html" title="Getting Started with Berkeley DB Transaction Processing" /> 10 <link rel="up" href="filemanagement.html" title="Chapter��5.��Managing DB Files" /> 11 <link rel="prev" href="filemanagement.html" title="Chapter��5.��Managing DB Files" /> 12 <link rel="next" href="recovery.html" title="Recovery Procedures" /> 13 </head> 14 <body> 15 <div class="navheader"> 16 <table width="100%" summary="Navigation header"> 17 <tr> 18 <th colspan="3" align="center">Backup Procedures</th> 19 </tr> 20 <tr> 21 <td width="20%" align="left"><a accesskey="p" href="filemanagement.html">Prev</a>��</td> 22 <th width="60%" align="center">Chapter��5.��Managing DB Files</th> 23 <td width="20%" align="right">��<a accesskey="n" href="recovery.html">Next</a></td> 24 </tr> 25 </table> 26 <hr /> 27 </div> 28 <div class="sect1" lang="en" xml:lang="en"> 29 <div class="titlepage"> 30 <div> 31 <div> 32 <h2 class="title" style="clear: both"><a id="backuprestore"></a>Backup Procedures</h2> 33 </div> 34 </div> 35 </div> 36 <div class="toc"> 37 <dl> 38 <dt> 39 <span class="sect2"> 40 <a href="backuprestore.html#copyutilities">About Unix Copy Utilities</a> 41 </span> 42 </dt> 43 <dt> 44 <span class="sect2"> 45 <a href="backuprestore.html#standardbackup">Offline Backups</a> 46 </span> 47 </dt> 48 <dt> 49 <span class="sect2"> 50 <a href="backuprestore.html#hotbackup">Hot Backup</a> 51 </span> 52 </dt> 53 <dt> 54 <span class="sect2"> 55 <a href="backuprestore.html#incrementalbackups">Incremental Backups</a> 56 </span> 57 </dt> 58 </dl> 59 </div> 60 <p> 61 <span class="emphasis"><em>Durability</em></span> is an important part of your 62 transactional guarantees. It means that once a transaction has been 63 successfully committed, your application will always see the results of that 64 transaction. 65 </p> 66 <p> 67 Of course, no software algorithm can guarantee durability in the face of physical data loss. Hard drives 68 can fail, and if you have not copied your data to locations other than your primary disk drives, 69 then you will lose data when those drives fail. Therefore, in order to truly obtain a durability 70 guarantee, you need to ensure that any data stored on disk is backed up to secondary or alternative storage, 71 such as secondary disk drives, or offline tapes. 72 </p> 73 <p> 74 There are three different types of backups that you can 75 perform with DB databases and log files. They are: 76 </p> 77 <div class="itemizedlist"> 78 <ul type="disc"> 79 <li> 80 <p> 81 Offline backups 82 </p> 83 <p> 84 This type of backup is perhaps the easiest to perform as it 85 involves simply copying database and log files to an 86 offline storage area. It also gives you a snapshot of the 87 database at a fixed, known point in time. However, you 88 cannot perform this type of a backup while you are performing 89 writes to the database. 90 </p> 91 </li> 92 <li> 93 <p> 94 Hot backups 95 </p> 96 <p> 97 This type of backup gives you a snapshot of your database. 98 Since your application can be writing to the database at the time that the 99 snapshot is being taken, you do not necessarily know what 100 the exact state of the database is for that given snapshot. 101 </p> 102 </li> 103 <li> 104 <p> 105 Incremental backups 106 </p> 107 <p> 108 This type of backup refreshes a previously performed backup. 109 </p> 110 </li> 111 </ul> 112 </div> 113 <p> 114 Once you have performed a backup, you can 115 perform <span class="emphasis"><em>catastrophic recovery</em></span> to restore 116 your databases from the backup. See 117 <a class="xref" href="recovery.html#catastrophicrecovery" title="Catastrophic Recovery">Catastrophic Recovery</a> 118 for more information. 119 </p> 120 <p> 121 Note that you can also maintain a hot failover. See 122 <a class="xref" href="hotfailover.html" title="Using Hot Failovers">Using Hot Failovers</a> 123 for more information. 124 </p> 125 <div class="sect2" lang="en" xml:lang="en"> 126 <div class="titlepage"> 127 <div> 128 <div> 129 <h3 class="title"><a id="copyutilities"></a>About Unix Copy Utilities</h3> 130 </div> 131 </div> 132 </div> 133 <p> 134 135 If you are copying database files you must copy databases atomically, 136 in multiples of the database page size. In other words, the reads made by 137 the copy program must not be interleaved with writes by 138 other threads of control, and the copy program must read the 139 databases in multiples of the underlying database page size. 140 Generally, this is not a problem because operating systems 141 already make this guarantee and system utilities normally 142 read in power-of-2 sized chunks, which are larger than the 143 largest possible Berkeley DB database page size. 144 </p> 145 <p> 146 On some platforms (most notably, some releases of Solaris), the copy utility (<code class="literal">cp</code>) was 147 implemented using the <code class="function">mmap()</code> system call rather than the 148 <code class="function">read()</code> system call. Because <code class="function">mmap()</code> did not make the same 149 guarantee of read atomicity as did <code class="function">read()</code>, the <code class="literal">cp</code> utility 150 could create corrupted copies of the databases. 151 </p> 152 <p> 153 Also, some platforms have implementations of the <code class="literal">tar</code> utility that performs 10KB block 154 reads by default. Even when an output block size is specified, the utility will still not read the 155 underlying databases in multiples of the specified block size. Again, the result can be a corrupted backup. 156 </p> 157 <p> 158 To fix these problems, use the <code class="literal">dd</code> utility instead of <code class="literal">cp</code> or 159 <code class="literal">tar</code>. When you use <code class="literal">dd</code>, make sure you specify a block size that is 160 equal to, or an even multiple of, your database page size. Finally, if you plan to use a system 161 utility to copy database files, you may want to use a system call trace utility (for example, 162 <code class="literal">ktrace</code> or <code class="literal">truss</code>) to make sure you are not using a I/O size that is 163 smaller than your database page size. You can also use these utilities to make sure the system utility is 164 not using a system call other than <code class="function">read()</code>. 165 </p> 166 </div> 167 <div class="sect2" lang="en" xml:lang="en"> 168 <div class="titlepage"> 169 <div> 170 <div> 171 <h3 class="title"><a id="standardbackup"></a>Offline Backups</h3> 172 </div> 173 </div> 174 </div> 175 <p> 176 To create an offline backup: 177 </p> 178 <div class="orderedlist"> 179 <ol type="1"> 180 <li> 181 <p> 182 Commit or abort all on-going transactions. 183 </p> 184 </li> 185 <li> 186 <p> 187 Pause all database writes. 188 </p> 189 </li> 190 <li> 191 <p> 192 Force a checkpoint. See 193 <a class="xref" href="filemanagement.html#checkpoints" title="Checkpoints">Checkpoints</a> 194 for details. 195 </p> 196 </li> 197 <li> 198 <p> 199 Copy all your database files to the backup location. 200 <span> 201 Note that you can simply copy all of the database 202 files, or you can determine which database files 203 have been written during the lifetime of the current 204 logs. To do this, use either the 205 206 <span> 207 <code class="methodname">Environment.getArchiveDatabases()</code>, 208 method 209 </span> 210 211 or use the <span class="command"><strong>db_archive</strong></span> 212 command with the <code class="literal">-s</code> option. 213 </span> 214 </p> 215 <p> 216 However, be aware that backing up just the modified databases only works if you have all of your 217 log files. If you have been removing log files for any reason then using 218 219 <span> 220 <code class="methodname">getArchiveDatabases()</code>, 221 </span> 222 223 can result in an 224 unrecoverable backup because you might not be notified of a database file that was modified. 225 </p> 226 </li> 227 <li> 228 <p> 229 Copy the <span class="emphasis"><em>last</em></span> log file to your backup location. 230 Your log files are named 231 <code class="literal">log.<span class="emphasis"><em>xxxxxxxxxx</em></span></code>, 232 where <span class="emphasis"><em>xxxxxxxxxx</em></span> is a 233 sequential number. The last log file is the file 234 with the highest number. 235 </p> 236 </li> 237 </ol> 238 </div> 239 </div> 240 <div class="sect2" lang="en" xml:lang="en"> 241 <div class="titlepage"> 242 <div> 243 <div> 244 <h3 class="title"><a id="hotbackup"></a>Hot Backup</h3> 245 </div> 246 </div> 247 </div> 248 <p> 249 To create a hot backup, you do not have to stop database 250 operations. Transactions may be on-going and you can be writing 251 to your database at the time of the backup. However, this means 252 that you do not know exactly what the state of your database is 253 at the time of the backup. 254 </p> 255 <p> 256 You can use the <span class="command"><strong>db_hotbackup</strong></span> command line utility to create a hot backup for you. This 257 utility will (optionally) run a checkpoint and the copy all necessary files to a target directory. 258 </p> 259 <p> 260 Alternatively, you can manually create a hot backup as follows: 261 </p> 262 <div class="orderedlist"> 263 <ol type="1"> 264 <li> 265 <p> 266 Copy all your database files to the backup location. 267 <span> 268 Note that you can simply copy all of the database 269 files, or you can determine which database files 270 have been written during the lifetime of the current 271 logs. To do this, use either the 272 273 <span> 274 <code class="methodname">Environment.getArchiveDatabases()</code>, 275 </span> 276 277 or use the <span class="command"><strong>db_archive</strong></span> 278 command with the <code class="literal">-s</code> option. 279 </span> 280 </p> 281 </li> 282 <li> 283 <p> 284 Copy all logs to your backup location. 285 </p> 286 </li> 287 </ol> 288 </div> 289 <div class="note" style="margin-left: 0.5in; margin-right: 0.5in;"> 290 <h3 class="title">Note</h3> 291 <p> 292 It is important to copy your database files <span class="emphasis"><em>and 293 then</em></span> your logs. In this way, 294 you can complete or roll back any database operations that were only partially completed 295 when you copied the databases. 296 </p> 297 </div> 298 </div> 299 <div class="sect2" lang="en" xml:lang="en"> 300 <div class="titlepage"> 301 <div> 302 <div> 303 <h3 class="title"><a id="incrementalbackups"></a>Incremental Backups</h3> 304 </div> 305 </div> 306 </div> 307 <p> 308 Once you have created a full backup (that is, either a 309 offline or hot backup), you can create incremental backups. 310 To do this, simply copy all of your currently existing log 311 files to your backup location. 312 </p> 313 <p> 314 Incremental backups do not require you to run a checkpoint 315 or to cease database write operations. 316 </p> 317 <p> 318 When you are working with incremental backups, remember 319 that the greater the number of log files contained in 320 your backup, the longer recovery will take. 321 You should run full backups 322 on some interval, and then do incremental backups on a shorter interval. 323 How frequently you need to run a full backup 324 is determined by the rate at which your databases change and 325 how sensitive your application is to lengthy recoveries 326 (should one be required). 327 </p> 328 <p> 329 You can also shorten recovery time by running recovery against the backup as you take each incremental 330 backup. Running recovery as you go means that there will be less work for DB to do if you should 331 ever need to restore your environment from the backup. 332 </p> 333 </div> 334 </div> 335 <div class="navfooter"> 336 <hr /> 337 <table width="100%" summary="Navigation footer"> 338 <tr> 339 <td width="40%" align="left"><a accesskey="p" href="filemanagement.html">Prev</a>��</td> 340 <td width="20%" align="center"> 341 <a accesskey="u" href="filemanagement.html">Up</a> 342 </td> 343 <td width="40%" align="right">��<a accesskey="n" href="recovery.html">Next</a></td> 344 </tr> 345 <tr> 346 <td width="40%" align="left" valign="top">Chapter��5.��Managing DB Files��</td> 347 <td width="20%" align="center"> 348 <a accesskey="h" href="index.html">Home</a> 349 </td> 350 <td width="40%" align="right" valign="top">��Recovery Procedures</td> 351 </tr> 352 </table> 353 </div> 354 </body> 355</html> 356