--- /dev/null
+<!-- doc/src/sgml/release-xl-10r1.sgml -->
+<!-- See header comment in release.sgml about typical markup -->
+ <sect1 id="release-xl-10r1beta1">
+ <title>Postgres-XL Release 10r1beta1</title>
+
+ <note>
+ <title>Release Date</title>
+ <simpara>2018-09-14</simpara>
+ <simpara>Current as of 2016-09-14</simpara>
+ </note>
+
+ <sect2>
+ <title>Overview</title>
+ <para>
+ This major release of <productname>Postgres-XL</productname> comes
+almost 2 years after <productname>Postgres-XL</productname> 9.5 was released,
+which was based on
+<productname>PostgreSQL</productname> 9.5. This release includes most of the
+new features added in <productname>PostgreSQL</productname> 9.6 and 10
+releases. This also includes almost all significant performance enhancements
+that were made to <productname>PostgreSQL</productname> in the last few years.
+ </para>
+
+ <para>
+ Apart from this, a lot of work has gone into fixing outstanding bugs,
+improving stability of the product and significantly improving performance for OLTP
+workloads. The following sections describe the major enhancements in
+<productname>PostgreSQL</productname> that are now available in
+<productname>Postgres-XL</productname>. A short list of features that are
+currently not supported or have limited support is also mentioned in
+<xref linkend="major-unsupported-enhancements">.
+ </para>
+ </sect2>
+
+ <sect2>
+ <title>Migration to Postgres-XL 10r1beta1 </title>
+ <sect3>
+ <title>Migration to Postgres-XL 10r1beta1 from PostgreSQL 10 </title>
+ <para>
+ A dump/restore using <xref linkend="app-pg-dumpall">
+ is required for those wishing to migrate
+ data from PostgreSQL 10 release. We don't currently support <xref
+linkend="pgupgrade">, as a mechanism to upgrade from PostgreSQL 10.
+ </para>
+
+ <para>
+ </para>
+ </sect3>
+ <sect3>
+ <title>Migration to Postgres-XL 10r1beta1 from Postgres-XL 9.5r1.x </title>
+ <para>
+ Version 10r1beta1 contains a number of changes that may affect compatibility
+ with previous releases. Since Postgres-XL 10r1beta1 release includes all changes
+from PostgreSQL 9.6 and PostgreSQL 10 releases, it is
+recommended that you review respective release notes of those releases to find
+the exact incompatibilities.
+ </para>
+ </sect3>
+ </sect2>
+
+ <sect2>
+ <title>Major Enhancements in Postgres-XL 10r1beta1 </title>
+ <para>
+ <productname>Postgres-XL</productname> 10r1beta1 is the first major release
+after <productname>Postgres-XL</productname> 9.5r1 release. So this release
+contains most of the major enhancements that went into
+<productname>PostgreSQL</productname> releases 9.6 and 10. This is a very
+short list of such enhancements, but all other enhancements also apply, unless
+otherwise stated.
+ </para>
+ <sect3>
+ <title>Major Enhancements from PostgreSQL 10 </title>
+ <itemizedlist>
+ <listitem><para>Declarative table partitioning</para></listitem>
+ <listitem><para>Improved query parallelism</para></listitem>
+ <listitem><para>Significant general performance improvements</para></listitem>
+ <listitem><para>Improved monitoring and control</para></listitem>
+ </itemizedlist>
+ </sect3>
+ <sect3>
+ <title>Major Enhancements from PostgreSQL 9.6 </title>
+ <itemizedlist>
+ <listitem><para>Parallel execution of sequential scans, joins and aggregates</para></listitem>
+ <listitem><para> Avoid scanning pages unnecessarily during vacuum freeze operations</para></listitem>
+ <listitem><para>Substantial performance improvements, especially in the area of scalability on multi-<acronym>CPU</acronym>-socket servers</para></listitem>
+ <listitem><para>Full-text search can now search for phrases (multiple adjacent words)</para></listitem>
+ </itemizedlist>
+ </sect3>
+ </sect2>
+
+ <sect2>
+ <title>Changes</title>
+
+ <para>
+ Below you will find a detailed account of the changes between
+ <productname>Postgres-XL</productname> 10r1beta1 and the previous major
+ release.
+ </para>
+
+ <sect3>
+ <title>Performance</title>
+ <itemizedlist>
+ <listitem>
+ <para>
+ Allow minimum value of pgxl_remote_fetch_size to zero
+ </para>
+ <para>
+ This is an experimental work. By setting pgxl_remote_fetch_size to 0, user can
+ fetch all rows at once from the remote side (instead of fetching a small number
+ at a time, thus giving the node to produce more rows as the previous set is
+ consumed). While fetching all rows at once is not very useful, it allows us to
+ circumvent PostgreSQL's limitation of not supporting parallel queries unless a
+ Portal is run once and to the end.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Automatically trigger ANALYZE (COORDINATOR) on remote coordinators
+ </para>
+ <para>
+ When a table is manually or automcatically analyzed, the coordiantor
+ will an ANALYZE (COORDINATOR) command to all other coordinators for them to
+ update their planner stats. This ensures that the planner stats remain up to date on all
+ coordinators in the cluster. The feature works when a table is analyzed
+ outside a transaction block.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Collect index statistics during ANALYZE on coordinator.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Randomize the choice of the initial ROUNDROBIN node.
+ </para>
+ </listitem>
+ </itemizedlist>
+ </sect3>
+ <sect3>
+ <title>Additional Features</title>
+ <itemizedlist>
+ <listitem>
+ <para>
+ <command>pgxc_ctl</command> is now a built-in command and gets
+automatically compiled and installed along with other binaries.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Enable Hot Standby on the replicas.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Add support for MODULO distribution on BIGINT.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Support an additional syntax ANALYZE (COORDINATOR) to allow users to rebuild
+ coordinator side statistics without running ANALYZE again on the datanodes.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Support GROUPING SETS, ROLLUP and CUBE.
+ </para>
+ </listitem>
+ </itemizedlist>
+ </sect3>
+ <sect3>
+ <title>Improvements to pgxc_ctl</title>
+ <itemizedlist>
+ <listitem>
+ <para>
+ <command>pgxc_ctl</command> is now a built-in command and gets
+automatically compiled and installed along with other binaries.
+ </para>
+ </listitem>
+ </itemizedlist>
+ </sect3>
+ <sect3>
+ <title>Misc Improvements</title>
+ <itemizedlist>
+ <listitem>
+ <para>
+ <command>pgxc_ctl</command> is now a built-in command and gets
+automatically compiled and installed along with other binaries.
+ </para>
+ </listitem>
+ </itemizedlist>
+ </sect3>
+ </sect2>
+ <sect2>
+ <title>Important Bug Fixes</title>
+ <itemizedlist>
+ <listitem>
+ <para>
+ Make improvements to cluster monitor process and global state
+reporting.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Correctly select the GTM proxy for a new node being added.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Fixes related to handling of schema qualification of temporary tables on remote nodes.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Fix handling of REFRESH MATERIALIZED VIEW CONCURRENTLY.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Improve locking semantics in GTM and GTM Proxy.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Ensure pooler process follows consistent model for SIGQUIT handling.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Fix broken implementation of recovery to barrier.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Track clearly whether to run a remote transaction in autocommit or a block.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Disable FQS for cursors defined with SCROLL.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Ensure that database objects are created consistently.
+ </para>
+
+ <para>
+ We now create views/materialised views on all nodes, unless they are temporary
+ objects in which case they are created only on the local coordinator and the
+ datanodes. Similarly, temporary sequences are created on the local coordinator
+ and the datanodes.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Prevent dropping distribution keys for MODULO.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Use thread-specific storage for computing a snapshot.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Add a user configurable parameter <xref linkend="guc-pgxl-remote-fetch-size"> to control the number of rows fetched from
+ the remote side during RemoteSubplan execution. Switch connections after processing pgxl_remote_fetch_size rows.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Allow COPY (INSERT RETURNING).
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Handle temp sequences so that duplicate values are not produced.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Ensure that the config parameters specified in the gtmPxyExtraConfig file take
+ precendence over the defaults.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Handle sequence's transactional behaviour on GTM.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Handle multi-command queries correctly inside SQL as well as plpgsql functions.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Do not silently skip FK constraints if loose_constraints are ON.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Ensure that the statistics about all-visible pages in fetched from remote
+ datanodes and updated at the coordinator for better planning.
+ </para>
+ </listitem>
+ </itemizedlist>
+ </sect2>
+ <sect2>
+ <title>Known Limitation</title>
+ <para>
+ While <productname>Postgres-XL</productname> strives hard to stay compatible
+with <productname>PostgreSQL</productname> so that applications running on
+<productname>PostgreSQL</productname> can be easily ported to
+<productname>Postgres-XL</productname>, there are certain known limitations of
+the product. Many of these can be addressed as and when sufficient development
+funding and resources become available.
+ <itemizedlist>
+ <listitem>
+ <para>
+ Distributed deadlock detection is not supported yet. Deadlock detection exists
+ on each node, just not across nodes.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Materialised views are currently maintained only on the coordinator.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ EXCLUSION CONSTRAINTS are enforced when both rows map to the same datanode.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ User defined functions have several limitations.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ There are restrictions on complex UPDATE/DELETE queries and updating
+distribution column values.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ TRIGGERs are not supported.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ EVENT TRIGGERs are not supported.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ SERIALIZABLE TRANSACTIONs are not supported.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ CREATE INDEX CONCURRENTLY is not supported.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ SAVEPOINTs are not supported.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Large objects are not supported.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Recursive queries work only in certain conditions.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Foreign Data Wrappers are not supported.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ INSENSITIVE/SCROLL/WITH HOLD cursors are not supported.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ LISTEN/NOTIFY is not supported.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ Remote Subplan nodes are marked parallel unsafe. This implies that
+ parallel workers can only be used for bottom most Remote Subplan node(s) in a
+ plan tree.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ CREATE PUBLICATION/SUBSCRIBER is not supported.
+ </para>
+ </listitem>
+ <listitem>
+ <para>
+ All partitions of a partitioned table must have consistent column ordering
+and column positioning.
+ </para>
+ </listitem>
+ </itemizedlist>
+ </para>
+ </sect2>
+ <sect2>
+ <title>Contributing to Postgres-XL</title>
+ <para>
+ Contributions to <productname>Postgres-XL</productname> are welcome. The
+code will be accepted under the same open source license that governs this
+released version of <productname>Postgres-XL</productname>. The authors of the
+patches will be credited appropriately in the release notes of the future
+releases.
+ </para>
+ </sect2>
+ <sect2>
+ <title>Credits</title>
+ <para>
+ <productname>Postgres-XL</productname> has been evolving over many years,
+originating with the <productname>GridSQL</productname>
+project and <productname>Postgres-XC</productname>, later combined as TransLattice Storm (StormDB).
+TransLattice open sourced the project, resulting in Postgres-XL 9.2. More
+recently, the EU-funded Big Data project AXLE funded the main work to bring the
+code up to date as <productname>Postgres-XL</productname> 9.5, allowing
+<productname>Postgres-XL</productname> to take advantage of the
+rapidly increasing BI features plugged into the core of
+<productname>PostgreSQL</productname>.
+ </para>
+ </sect2>
+</sect1>
projects / postgres-xl.git / commitdiff