<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:taxo="http://purl.org/rss/1.0/modules/taxonomy/" version="2.0">
  <channel>
    <title>topic Index Tip Recovery For Large Repositories in Alfresco Archive</title>
    <link>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157728#M111892</link>
    <description>&lt;HTML&gt;&lt;HEAD&gt;&lt;/HEAD&gt;&lt;BODY&gt;&lt;SPAN&gt;Hi All,&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;After bashing my head against a wall trying to figure out how to recover our indexes with 300 000 transaction and having it take well over 24 hours I have put together some helpful tips to help any poor souls in a similar situation.&amp;nbsp; This was done against 2.1 CE.&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;PRE class="language-none line-numbers"&gt;&lt;CODE&gt;node.index.FullIndexRecoveryComponent&lt;SPAN class="line-numbers-rows"&gt;&lt;SPAN&gt;‍&lt;/SPAN&gt;&lt;/SPAN&gt;&lt;/CODE&gt;&lt;/PRE&gt;&lt;BR /&gt;&lt;SPAN&gt;- Firstly remember to blow away the existing &lt;/SPAN&gt;&lt;PRE class="language-none line-numbers"&gt;&lt;CODE&gt;lucene-indexes&lt;SPAN class="line-numbers-rows"&gt;&lt;SPAN&gt;‍&lt;/SPAN&gt;&lt;/SPAN&gt;&lt;/CODE&gt;&lt;/PRE&gt;&lt;SPAN&gt; directory AFTER shutting down tomcat.&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;- Set &lt;/SPAN&gt;&lt;PRE class="language-none line-numbers"&gt;&lt;CODE&gt;index.recovery.mode=FULL&lt;SPAN class="line-numbers-rows"&gt;&lt;SPAN&gt;‍&lt;/SPAN&gt;&lt;/SPAN&gt;&lt;/CODE&gt;&lt;/PRE&gt;&lt;BR /&gt;&lt;SPAN&gt;- One of the most frustrating things is that it is really hard to figure out if something is actually happening.&amp;nbsp; To make your life easier edit your log4j.properties (under FC with WAR install in /var/lib/tomcat5/webapps/alfresco/WEB-INF/classes/log4j.properties) and set&amp;nbsp; &lt;/SPAN&gt;&lt;PRE class="language-none line-numbers"&gt;&lt;CODE&gt;log4j.logger.org.alfresco.repo.node.index.FullIndexRecoveryComponent=debug&lt;SPAN class="line-numbers-rows"&gt;&lt;SPAN&gt;‍&lt;/SPAN&gt;&lt;/SPAN&gt;&lt;/CODE&gt;&lt;/PRE&gt;&lt;BR /&gt;&lt;SPAN&gt;You should now get &lt;/SPAN&gt;&lt;PRE class="language-none line-numbers"&gt;&lt;CODE&gt;Reindexing transaction: xxxxx&lt;SPAN class="line-numbers-rows"&gt;&lt;SPAN&gt;‍&lt;/SPAN&gt;&lt;/SPAN&gt;&lt;/CODE&gt;&lt;/PRE&gt;&lt;SPAN&gt; messages which will help you figure out what is going on&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;- I changed &lt;/SPAN&gt;&lt;PRE class="language-none line-numbers"&gt;&lt;CODE&gt;lucene.indexer.batchSize=1000&lt;SPAN class="line-numbers-rows"&gt;&lt;SPAN&gt;‍&lt;/SPAN&gt;&lt;/SPAN&gt;&lt;/CODE&gt;&lt;/PRE&gt;&lt;SPAN&gt; from its really high value.&amp;nbsp; This breaks the index recovery into chunks and has the happy effect of flushing to disk the lucene indexes to the lucene-indexes directory.&amp;nbsp; This flushing makes the log percentage indicator more accurate and seems to prevent nasty risks like the disk running out of space or the JVM running out of heap from being oh so heart breaking&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;- As per &lt;/SPAN&gt;&lt;A href="http://www.onjava.com/pub/a/onjava/2003/03/05/lucene.html" rel="nofollow noopener noreferrer"&gt;http://www.onjava.com/pub/a/onjava/2003/03/05/lucene.html&lt;/A&gt;&lt;SPAN&gt; benchmarks and recommendation for batch benchmarking I jacked the &lt;/SPAN&gt;&lt;BR /&gt;&lt;PRE class="language-none line-numbers"&gt;&lt;CODE&gt;lucene.indexer.mergeFactor=1000&lt;SPAN class="line-numbers-rows"&gt;&lt;SPAN&gt;‍&lt;/SPAN&gt;&lt;/SPAN&gt;&lt;/CODE&gt;&lt;/PRE&gt;&lt;SPAN&gt; from 10 .&amp;nbsp; However I did not benchmark this change in any rigorous fashion.&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;- After it is done you should set the batch size back as there is an open issue AR-1280 (documented in the file) about something bad happening if you don't&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;- Also remember to set the &lt;/SPAN&gt;&lt;PRE class="language-none line-numbers"&gt;&lt;CODE&gt;index.recovery.mode=AUTO&lt;SPAN class="line-numbers-rows"&gt;&lt;SPAN&gt;‍&lt;/SPAN&gt;&lt;/SPAN&gt;&lt;/CODE&gt;&lt;/PRE&gt;&lt;SPAN&gt; instead of FULL.&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;Cheers!&lt;/SPAN&gt;&lt;/BODY&gt;&lt;/HTML&gt;</description>
    <pubDate>Tue, 12 Feb 2008 02:58:43 GMT</pubDate>
    <dc:creator>callermd</dc:creator>
    <dc:date>2008-02-12T02:58:43Z</dc:date>
    <item>
      <title>Index Tip Recovery For Large Repositories</title>
      <link>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157728#M111892</link>
      <description>Hi All,After bashing my head against a wall trying to figure out how to recover our indexes with 300 000 transaction and having it take well over 24 hours I have put together some helpful tips to help any poor souls in a similar situation.&amp;nbsp; This was done against 2.1 CE.node.index.FullIndexRecoveryCo</description>
      <pubDate>Tue, 12 Feb 2008 02:58:43 GMT</pubDate>
      <guid>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157728#M111892</guid>
      <dc:creator>callermd</dc:creator>
      <dc:date>2008-02-12T02:58:43Z</dc:date>
    </item>
    <item>
      <title>Re: Index Tip Recovery For Large Repositories</title>
      <link>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157729#M111893</link>
      <description>&lt;HTML&gt;&lt;HEAD&gt;&lt;/HEAD&gt;&lt;BODY&gt;&lt;SPAN&gt;Hi&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;I have rasied an enhancement request to merge many small transactions together during index recovery.&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;The merge factor is not likely to make much difference for small transactions.&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;Moving to lucene 2.3 is likely to give significantly better performance.&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;Your perfromance seems very slow. What hardware are you using?&lt;/SPAN&gt;&lt;BR /&gt;&lt;SPAN&gt;Where is the alf data stored? What DB? Have you configured to do most content transforms atomically - or only "short" ones (the default). Roughly how many files/folders per transaction?&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;Andy&lt;/SPAN&gt;&lt;/BODY&gt;&lt;/HTML&gt;</description>
      <pubDate>Wed, 20 Feb 2008 12:51:38 GMT</pubDate>
      <guid>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157729#M111893</guid>
      <dc:creator>andy</dc:creator>
      <dc:date>2008-02-20T12:51:38Z</dc:date>
    </item>
    <item>
      <title>Re: Index Tip Recovery For Large Repositories</title>
      <link>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157730#M111894</link>
      <description>&lt;HTML&gt;&lt;HEAD&gt;&lt;/HEAD&gt;&lt;BODY&gt;&lt;SPAN&gt;Hi callermd,&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;This is an extremely valuable post, thanks! We recently rebuild our indexes (139,071 transactions) and that takes 10 hours on our server so next time we'll try your solution, 10 hours downtime for something like this is not really acceptable. It would be nice if Alfresco could tune this for the following release, hint, hint.&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;Can you tell us what the performance gain was after your tweaks? How much time did you save after this?&lt;/SPAN&gt;&lt;/BODY&gt;&lt;/HTML&gt;</description>
      <pubDate>Fri, 21 Mar 2008 09:14:09 GMT</pubDate>
      <guid>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157730#M111894</guid>
      <dc:creator>simon</dc:creator>
      <dc:date>2008-03-21T09:14:09Z</dc:date>
    </item>
    <item>
      <title>Re: Index Tip Recovery For Large Repositories</title>
      <link>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157731#M111895</link>
      <description>&lt;HTML&gt;&lt;HEAD&gt;&lt;/HEAD&gt;&lt;BODY&gt;&lt;SPAN&gt;Hi&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;The index rebuild is now multi-threaded. &lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;Andy&lt;/SPAN&gt;&lt;/BODY&gt;&lt;/HTML&gt;</description>
      <pubDate>Tue, 15 Jul 2008 10:17:52 GMT</pubDate>
      <guid>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157731#M111895</guid>
      <dc:creator>andy</dc:creator>
      <dc:date>2008-07-15T10:17:52Z</dc:date>
    </item>
    <item>
      <title>Re: Index Tip Recovery For Large Repositories</title>
      <link>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157732#M111896</link>
      <description>&lt;HTML&gt;&lt;HEAD&gt;&lt;/HEAD&gt;&lt;BODY&gt;&lt;BLOCKQUOTE class="jive-quote"&gt;Hi&lt;BR /&gt;&lt;BR /&gt;The index rebuild is now multi-threaded. &lt;BR /&gt;&lt;BR /&gt;Andy&lt;/BLOCKQUOTE&gt;&lt;BR /&gt;&lt;SPAN&gt;What version will have this functionality?&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;Thanks!&lt;/SPAN&gt;&lt;/BODY&gt;&lt;/HTML&gt;</description>
      <pubDate>Tue, 05 Aug 2008 21:13:23 GMT</pubDate>
      <guid>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157732#M111896</guid>
      <dc:creator>callermd</dc:creator>
      <dc:date>2008-08-05T21:13:23Z</dc:date>
    </item>
    <item>
      <title>Re: Index Tip Recovery For Large Repositories</title>
      <link>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157733#M111897</link>
      <description>&lt;HTML&gt;&lt;HEAD&gt;&lt;/HEAD&gt;&lt;BODY&gt;&lt;SPAN&gt;Putting a backup process in place (as described at &lt;/SPAN&gt;&lt;A href="http://wiki.alfresco.com/wiki/Backup_and_Restore" rel="nofollow noopener noreferrer"&gt;http://wiki.alfresco.com/wiki/Backup_and_Restore&lt;/A&gt;&lt;SPAN&gt;) can help avoid these problems.&amp;nbsp; See &lt;/SPAN&gt;&lt;A href="http://wiki.alfresco.com/wiki/Backup_and_Restore#Lucene_Index_Restoration" rel="nofollow noopener noreferrer"&gt;http://wiki.alfresco.com/wiki/Backup_and_Restore#Lucene_Index_Restoration&lt;/A&gt;&lt;SPAN&gt; specifically.&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;Cheers,&lt;/SPAN&gt;&lt;BR /&gt;&lt;SPAN&gt;Peter&lt;/SPAN&gt;&lt;/BODY&gt;&lt;/HTML&gt;</description>
      <pubDate>Tue, 05 Aug 2008 23:43:16 GMT</pubDate>
      <guid>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157733#M111897</guid>
      <dc:creator>pmonks</dc:creator>
      <dc:date>2008-08-05T23:43:16Z</dc:date>
    </item>
    <item>
      <title>Re: Index Tip Recovery For Large Repositories</title>
      <link>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157734#M111898</link>
      <description>&lt;HTML&gt;&lt;HEAD&gt;&lt;/HEAD&gt;&lt;BODY&gt;&lt;SPAN&gt;You should probably add to that:&lt;/SPAN&gt;&lt;BR /&gt;&lt;A href="http://forums.alfresco.com/en/viewtopic.php?f=4&amp;amp;t=15401" rel="nofollow noopener noreferrer"&gt;http://forums.alfresco.com/en/viewtopic.php?f=4&amp;amp;t=15401&lt;/A&gt;&lt;BR /&gt;&lt;SPAN&gt;We are running on 5-10m transactions…&lt;/SPAN&gt;&lt;BR /&gt;&lt;BR /&gt;&lt;SPAN&gt;Oz&lt;/SPAN&gt;&lt;/BODY&gt;&lt;/HTML&gt;</description>
      <pubDate>Sat, 04 Jul 2009 01:47:44 GMT</pubDate>
      <guid>https://connect.hyland.com/t5/alfresco-archive/index-tip-recovery-for-large-repositories/m-p/157734#M111898</guid>
      <dc:creator>oznevo</dc:creator>
      <dc:date>2009-07-04T01:47:44Z</dc:date>
    </item>
  </channel>
</rss>

