[teiid-commits] teiid SVN: r2979 - trunk/documentation/admin-guide/src/main/docbook/en-US/content.
teiid-commits at lists.jboss.org
teiid-commits at lists.jboss.org
Tue Mar 8 15:47:01 EST 2011
Author: shawkins
Date: 2011-03-08 15:47:01 -0500 (Tue, 08 Mar 2011)
New Revision: 2979
Modified:
trunk/documentation/admin-guide/src/main/docbook/en-US/content/performance.xml
Log:
TEIID-1502 update to the performance doc to explain the thread settings and more on the max reserve columns
Modified: trunk/documentation/admin-guide/src/main/docbook/en-US/content/performance.xml
===================================================================
--- trunk/documentation/admin-guide/src/main/docbook/en-US/content/performance.xml 2011-03-08 20:39:59 UTC (rev 2978)
+++ trunk/documentation/admin-guide/src/main/docbook/en-US/content/performance.xml 2011-03-08 20:47:01 UTC (rev 2979)
@@ -35,6 +35,15 @@
If you are processing large (100s of MBs) of highly unique datasets
through Teiid, you should consider <link linkend="system_properties">disabling value caching</link> since it
will not significantly reduce memory consumption.
+ <note>
+ <para>The formula (average bytes per value)*(processor batch size)*(maxReserveBatchColumns) approximates the amount of
+ reserve memory used by the BufferManager. The defaults with an assumed value (8 bytes in 32 bit mode) for bytes per column
+ works out to be: 8 bytes * 512 * 16384 = 64MB. Memory consumption can be significantly more or less
+ depending upon actual column values and whether value caching is enabled. The 8 bytes per column includes row/batch list overhead and
+ represents the Java type memory foot print (e.g. java.lang.Integer not a raw 4 byte integer).
+ The nominal target of 64MB per 1 gigabyte of may be too low based upon the particulars of your usage. Raising the maxReserveBatchColumns
+ value too high though may result in out of memory errors.</para>
+ </note>
</para>
<para>
Each intermediate result buffer, temporary LOB, and temporary table
@@ -78,11 +87,15 @@
JDBC, ODBC, and Admin access.
Typical installations will not need to
adjust the default thread and buffer
- size settings. At this time, ODBC
+ size settings.
+ All JDBC socket operations are non-blocking, so setting the number of maxThreads
+ higher than the maximum effective parallelism of the machine will not result in greater performance. The default value 0 for JDBC socket threads will
+ set the max to the number of available processors.
+ At this time, ODBC
queries are executed synchronously from the socket thread.
Simultaneous long-running queries may exhaust the available threads.
Consider increasing the default max threads (15) for ODBC if you
- expect a higher concurrent load of long-running queries.</para>
+ expect a higher concurrent load of long-running ODBC client queries.</para>
</section>
<section>
<title>LOBs</title>
More information about the teiid-commits
mailing list