Currently the package contains the only tool pgcompact
, we are
planning to add much more in the future. Stay tuned.
The list of changes can be found in CHANGES.md. The To-Do List is in TODO.md.
The easiest way to deploy the toolkit is to download the latest
stable version, unpack the archive and copy the autonomous scripts
from the fatpack
directory to your bin
or to just add this
directory to your PATH
. The autonomous scripts are packed with all
the dependencies and need Perl >=5.8.8
to be installed only.
If you need a specific version or branch of the tools, replace the
stable
with its name string in the URL.
Another way is to git clone
the repository
git clone https://github.com/grayhemp/pgtoolkit.git
or to svn checkout
it
svn checkout https://github.com/grayhemp/pgtoolkit
Do not forget to switch to the necessary version branch afterwards.
You can also use the autonomous scripts in the fatpack
directory or
the non-autonomous versions of them in the bin
directory. For the
latter you need the lib
directory either to be in the same
sub-directory with bin
or to be in your PERL5LIB
.
A tool to reduce bloat for tables and indexes without heavy locks and full table rebuilding.
Initially the tool is an automation of the solutions proposed in these publications:
- Reducing bloat without locking by Joshua Tolley
- Reduce bloat of table without long/exclusive locks by Hubert Lubaczewski.
If pgstattuple is installed pgcompact
uses it to get a better
statistics. It is highly recommended to be for TOAST
ed tables and
indexes.
Shows user manual.
pgcompact --man
Compacts all the bloated tables in all the databases in the cluster plus their bloated indexes. Prints additional progress information.
pgcompact --all --reindex --verbosity info
Compacts all the bloated tables in the billing database and their
bloated indexes excepts ones that are in the pgq
schema.
pgcompact --dbname billing --exclude-schema pgq --reindex
- Requires no dependencies except
Perl >=5.8.8
, so it can just be copied to server and run - Works via
DBD::Pg
,DBD::PgPP
or evenpsql
if there are no former ones, detects and chooses the best option automatically - Processes either whole cluster or specified tables, schemes, databases only
- Has an ability to exclude tables, schemes or databases from processing
- Performs bloat analysis and processes those tables that have it only. We recommend to install pgstattuple for more precise estimations.
- Uses non blocking reindex techniques
- Performs indexes bloat analysis and processes only the required ones
- Analyses and rebuilds bloated unique constraints and primary keys where possible
- Provides TOAST tables and their indexes bloat information and rebuilding instructions
- Incremental processing, in other words one can stop the process and continue it at any time later
- Dynamically adjusts behavior for current load of database to not affect its performance
- Can be run in several parallel sessions on the same instance to process the tables faster
- Instructs administrators, supplying them with ready to use DDL, to manually rebuild database objects that can not be rebuilt automatically
- PgCookbook - a PostgreSQL documentation project
Copyright © 2011-2014 Sergey Konoplev, Maxim Boguk
PgToolkit is released under the PostgreSQL License, read LICENSE.md for additional information.
Thank you:
- DenisBY for bug reports and testing
- PostgreSQL-Consulting.com for a huge amount of ideas and lots of testing
- Maxim Boguk for ideas, testing and useful hints
- Lonni Friedman for your ideas
- Rural Hunter for ideas and testing
- Hubert "depesz" Lubaczewski for testing, useful hints and code contributions
- Gonzalo Gil for testing.