| List: | Commits | « Previous MessageNext Message » | |
| From: | jon | Date: | March 27 2008 5:47pm |
| Subject: | svn commit - mysqldoc@docsrva: r10353 - in trunk: refman-4.1 refman-5.0 refman-5.1 | ||
| View as plain text | |||
Author: jstephens Date: 2008-03-27 17:47:46 +0100 (Thu, 27 Mar 2008) New Revision: 10353 Log: "18.75% + 1" overhead for DataMemory (Thanks, Jonas and ToddF!) Modified: trunk/refman-4.1/mysql-cluster-configuration.xml trunk/refman-5.0/mysql-cluster-configuration.xml trunk/refman-5.1/mysql-cluster-configuration.xml Modified: trunk/refman-4.1/mysql-cluster-configuration.xml =================================================================== --- trunk/refman-4.1/mysql-cluster-configuration.xml 2008-03-27 14:41:40 UTC (rev 10352) +++ trunk/refman-4.1/mysql-cluster-configuration.xml 2008-03-27 16:47:46 UTC (rev 10353) Changed blocks: 1, Lines Added: 24, Lines Deleted: 0; 1808 bytes @@ -1603,6 +1603,30 @@ </para> <para> + In addition, due to the way in which new pages are allocated + when the capacity of the current page is exhausted, there is + an additional overhead of approximately 18.75%. When more + <literal>DataMemory</literal> is required, more than one new + page is allocated, according to the following formula: + +<programlisting> +number of new pages = FLOOR(number of current pages × 0.1875) + 1 +</programlisting> + + For example, if 15 pages are currently allocated to a given + table and an insert to this table requires additional + storage space, the number of new pages allocated to the + table is <literal>FLOOR(15 × 0.1875) + 1 = + FLOOR(2.8125) + 1 = 2 + 1 = + 3</literal>. Now 15 + 3 = 18 memory pages are + allocated to the table. When the last of these 18 pages + becomes full, <literal>FLOOR(18 × 0.1875) + 1 + = FLOOR(3.3750) + 1 = 3 + 1 = + 4</literal> new pages are allocated, so the total number of + pages allocated to the table is now 22. + </para> + + <para> Once a page has been allocated, it is currently not possible to return it to the pool of free pages, except by deleting the table. (This also means that Modified: trunk/refman-5.0/mysql-cluster-configuration.xml =================================================================== --- trunk/refman-5.0/mysql-cluster-configuration.xml 2008-03-27 14:41:40 UTC (rev 10352) +++ trunk/refman-5.0/mysql-cluster-configuration.xml 2008-03-27 16:47:46 UTC (rev 10353) Changed blocks: 1, Lines Added: 24, Lines Deleted: 0; 1808 bytes @@ -1580,6 +1580,30 @@ </para> <para> + In addition, due to the way in which new pages are allocated + when the capacity of the current page is exhausted, there is + an additional overhead of approximately 18.75%. When more + <literal>DataMemory</literal> is required, more than one new + page is allocated, according to the following formula: + +<programlisting> +number of new pages = FLOOR(number of current pages × 0.1875) + 1 +</programlisting> + + For example, if 15 pages are currently allocated to a given + table and an insert to this table requires additional + storage space, the number of new pages allocated to the + table is <literal>FLOOR(15 × 0.1875) + 1 = + FLOOR(2.8125) + 1 = 2 + 1 = + 3</literal>. Now 15 + 3 = 18 memory pages are + allocated to the table. When the last of these 18 pages + becomes full, <literal>FLOOR(18 × 0.1875) + 1 + = FLOOR(3.3750) + 1 = 3 + 1 = + 4</literal> new pages are allocated, so the total number of + pages allocated to the table is now 22. + </para> + + <para> Once a page has been allocated, it is currently not possible to return it to the pool of free pages, except by deleting the table. (This also means that Modified: trunk/refman-5.1/mysql-cluster-configuration.xml =================================================================== --- trunk/refman-5.1/mysql-cluster-configuration.xml 2008-03-27 14:41:40 UTC (rev 10352) +++ trunk/refman-5.1/mysql-cluster-configuration.xml 2008-03-27 16:47:46 UTC (rev 10353) Changed blocks: 1, Lines Added: 37, Lines Deleted: 0; 2205 bytes @@ -1654,6 +1654,43 @@ </para> <para> + In addition, due to the way in which new pages are allocated + when the capacity of the current page is exhausted, there is + an additional overhead of approximately 18.75%. When more + <literal>DataMemory</literal> is required, more than one new + page is allocated, according to the following formula: + +<programlisting> +number of new pages = FLOOR(number of current pages × 0.1875) + 1 +</programlisting> + + For example, if 15 pages are currently allocated to a given + table and an insert to this table requires additional + storage space, the number of new pages allocated to the + table is <literal>FLOOR(15 × 0.1875) + 1 = + FLOOR(2.8125) + 1 = 2 + 1 = + 3</literal>. Now 15 + 3 = 18 memory pages are + allocated to the table. When the last of these 18 pages + becomes full, <literal>FLOOR(18 × 0.1875) + 1 + = FLOOR(3.3750) + 1 = 3 + 1 = + 4</literal> new pages are allocated, so the total number of + pages allocated to the table is now 22. + </para> + + &mccge-warning-begin; + + <note> + <para> + The <quote>18.75% + 1</quote> overhead is no longer + required in &mccge-series; + 6.2.<replaceable>x</replaceable> and + 6.3.<replaceable>x</replaceable> releases. + </para> + </note> + + &mccge-warning-end-cluster; + + <para> Once a page has been allocated, it is currently not possible to return it to the pool of free pages, except by deleting the table. (This also means that
| Thread | ||
|---|---|---|
| • svn commit - mysqldoc@docsrva: r10353 - in trunk: refman-4.1 refman-5.0 refman-5.1 | jon | 27 Mar |
© 1995-2008 MySQL AB, 2008-2009 Sun Microsystems, Inc.
Page generated in 0.100 sec. using MySQL 5.1.14-beta-log