提交 e3ac474e 编写于 作者: M mkiyama 提交者: dyozie

GPDB 4.3.14.0 - clarify GUC max_appendonly_tables [#145961475]

上级 27d0ae5d
......@@ -6422,10 +6422,20 @@
<topic id="max_appendonly_tables">
<title>max_appendonly_tables</title>
<body>
<p>Sets the maximum number of append-optimized relations that can be written to or loaded
concurrently. Append-optimized table partitions and subpartitions are considered as unique
tables against this limit. Increasing the limit will allocate more shared memory at server
start.</p>
<p>Sets the maximum number of concurrent transactions that can write to or update
append-optimized tables. Transactions that exceed the maximum return an error.</p>
<p>Operations that are counted are <codeph>INSERT</codeph>, <codeph>UPDATE</codeph>,
<codeph>COPY</codeph>, and <codeph>VACUUM</codeph> operations. The limit is only for
in-progress transactions. Once a transaction ends (either aborted or committed), it is no
longer counted against this limit.</p>
<p>For operations against a partitioned table, each subpartition (child table) that is an
append-optimized table and is changed counts as a single table towards the maximum. For
example, a partitioned table <codeph>p_tbl</codeph> is defined with three subpartitions that
are append-optimized tables <codeph>p_tbl_ao1</codeph>, <codeph>p_tbl_ao2</codeph>, and
<codeph>p_tbl_ao3</codeph>. An <codeph>INSERT</codeph> or <codeph>UPDATE</codeph> command
against the partitioned table <codeph>p_tbl</codeph> that changes append-optimized tables
<codeph>p_tbl_ao1</codeph> and <codeph>p_tbl_ao2</codeph> is counted as two transactions. </p>
<p>Increasing the limit allocates more shared memory on the master host at server start.</p>
<table id="max_appendonly_tables_table">
<tgroup cols="3">
<colspec colnum="1" colname="col1" colwidth="1*"/>
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册