mirrors/postgres
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Add descriptions to TODO items and make adjustments based on 7.5.

Browse Source
This commit is contained in:
Bruce Momjian 2004-08-01 05:15:58 +00:00
parent 6e5e78241b
commit 5b0d5ea92d
1 changed files with 443 additions and 170 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

609
doc/TODO
View File

@ -5,10 +5,11 @@ TODO list for PostgreSQL
Bracketed items "[]" have more detail.
Current maintainer: Bruce Momjian (pgman@candle.pha.pa.us)
Last updated: Sat Jul 31 02:13:51 EDT 2004
Last updated: Sun Aug 1 01:15:12 EDT 2004
The most recent version of this document can be viewed at the PostgreSQL web site, http://www.PostgreSQL.org.
Remove items before beta?
Urgent
======
@ -20,7 +21,7 @@ Urgent
Administration
==============
* Incremental backups
* -Incremental backups
* Remove behavior of postmaster -o after making postmaster/postgres
flags unique
* -Allow configuration files to be specified in a different directory
@ -31,34 +32,73 @@ Administration
* -Allow logging of only data definition(DDL), or DDL and modification statements
* -Allow log lines to include session-level information, like database and user
* Allow server log information to be output as INSERT statements
This would allow server log information to be easily loaded into
a database for analysis.
* Prevent default re-use of sysids for dropped users and groups
Currently, if a user is removed while he still owns objects, a new
user given might be given their user id and inherit the
previous users objects.
* Prevent dropping user that still owns objects, or auto-drop the objects
* Allow pooled connections to query prepared queries
* Allow pooled connections to close all open WITH HOLD cursors
* Allow pooled connections to list all prepared queries
This would allow an application inheriting a pooled connection to know
the queries prepared in the current session.
* Allow major upgrades without dump/reload, perhaps using pg_upgrade
* Have SHOW ALL and pg_settings show descriptions for server-side variables(Joe)
* Allow external interfaces to extend the GUC variable set
* Allow GRANT/REVOKE permissions to be given to all schema objects with one command
* Have SHOW ALL and pg_settings show descriptions for server-side variables
* -Allow external interfaces to extend the GUC variable set
* Allow GRANT/REVOKE permissions to be given to all schema objects with one
command
* Remove unreferenced table files created by transactions that were
in-progress when the server terminated abruptly
* Allow reporting of which objects are in which tablespaces
This item is difficult because a tablespace can contain objects from
multiple databases. There is a server-side function that returns the
databases which use a specific tablespace, so this requires a tool
that will call that function and connect to each database to find the
objects in each database for that tablespace.
* Allow database recovery where tablespaces can't be created
* Add include functionality to postgresql.conf
* Allow changing of already-created database and schema tablespaces
* Allow moving system tables to other tablespaces, where possible
When a pg_dump is restored, all tablespaces will attempt to be created
in their original locations. If this fails, the user must be able to
adjust the restore process.
* Add "include file" functionality in postgresql.conf
* Add session start time and last statement time to pg_stat_activity
* Allow server logs to be read using SQL commands
* Allow server configuration parameters to be modified remotetly
* Allow server logs to be remotely read using SQL commands
* Allow server configuration parameters to be remotely modified
* Allow administrators to safely terminate individual sessions
* Allow point-in-time recovery to archive partially filled logs
Right now, SIGTERM will terminate a session, but it is treated as
though the postmaster has paniced and shared memory might not be
cleaned up properly. A new signal is needed for safe termination.
* Allow point-in-time recovery to archive partially filled write-ahead
logs
Currently only full WAL files are archived. This means that the most
recent transactions aren't available for recovery in case of a disk
failure.
* Improve replication solutions
o Automatic failover
The proper solution to this will probably the use of a master/slave
replication solution like Sloney and a connection pooling tool like
pgpool.
o Load balancing
o Master/slave replication
o Multi-master replication
o Partition data across servers
o Queries across databases or servers (two-phase commit)
You can use any of the master/slave replication servers to use a
standby server for data warehousing. To allow read/write queries to
multiple servers, you need multi-master replication like pgcluster.
o Allow replication over unreliable or non-persistent links
@ -68,24 +108,29 @@ Data Types
* Remove Money type, add money formatting for decimal type
* -Change factorial to return a numeric (Gavin)
* Change NUMERIC to enforce the maximum precision, and increase it
* Add function to return compressed length of TOAST data values (Tom)
* Allow INET subnet tests using non-constants to be indexed
* Add function to return compressed length of TOAST data values
* Allow INET subnet tests with non-constants to be indexed
* Add transaction_timestamp(), statement_timestamp(), clock_timestamp() functionality
* Have sequence dependency track use of DEFAULT sequences, seqname.nextval
* Disallow changing default expression of a SERIAL column
Current CURRENT_TIMESTAMP returns the start time of the current
transaction, and gettimeofday() returns the wallclock time. This will
make time reporting more consistent and will allow reporting of
the statement start time.
* Have sequence dependency track use of DEFAULT sequences,
seqname.nextval (?)
* Disallow changing default expression of a SERIAL column (?)
* Allow infinite dates just like infinite timestamps
* -Allow pg_dump to dump sequences using NO_MAXVALUE and NO_MINVALUE
* Allow backend to output result sets in XML
* -Prevent whole-row references from leaking memory, e.g. SELECT COUNT(tab.*)
* Have initdb set DateStyle based on locale?
* Add pg_get_acldef(), pg_get_typedefault(), and pg_get_attrdef()
* Add ALTER DOMAIN, AGGREGATE, CONVERSION, SEQUENCE ... OWNER TO
* Allow to_char to print localized month names (Karel)
* Allow to_char to print localized month names
* Allow functions to have a search path specified at creation time
* -Make LENGTH() of CHAR() not count trailing spaces
* Allow substring/replace() to get/set bit values
* Add GUC variable to allow output of interval values in ISO8601 format
* Support composite types as table columns
* -Support composite types as table columns
* Fix data types where equality comparison isn't intuitive, e.g. box
@ -93,31 +138,43 @@ Data Types
o Allow nulls in arrays
o Allow MIN()/MAX() on arrays
o Delay resolution of array expression type so assignment coercion
can be performed on empty array expressions (Joe)
can be performed on empty array expressions
o Modify array literal representation to handle array index lower bound
of other than one
* BINARY DATA
o Improve vacuum of large objects, like /contrib/vacuumlo
o Improve vacuum of large objects, like /contrib/vacuumlo (?)
o Add security checking for large objects
o Make file in/out interface for TOAST columns, similar to large object
interface (force out-of-line storage and no compression)
Currently large objects entries do not have owners. Permissions can
only be set at the pg_largeobject table level.
o Auto-delete large objects when referencing row is deleted
o Allow read/write into TOAST values like large objects
This requires the TOAST column to be stored EXTERNAL.
Multi-Language Support
======================
* Add NCHAR (as distinguished from ordinary varchar),
* Allow locale to be set at database creation
* Allow locale on a per-column basis, default to ASCII
* Optimize locale to have minimal performance impact when not used (Peter E)
Currently locale can only be set during initdb.
* Allow encoding on a per-column basis
Right now only one encoding is allowed per database.
* Optimize locale to have minimal performance impact when not used
* Support multiple simultaneous character sets, per SQL92
* Improve Unicode combined character handling
* Add octet_length_server() and octet_length_client() (Thomas, Tatsuo)
* Make octet_length_client the same as octet_length() (?)
* Prevent mismatch of frontend/backend encodings from converting bytea
* Improve Unicode combined character handling (?)
* Add octet_length_server() and octet_length_client()
* Make octet_length_client() the same as octet_length()?
* -Prevent mismatch of frontend/backend encodings from converting bytea
data from being interpreted as encoded strings
* -Fix upper()/lower() to work for multibyte encodings
@ -136,69 +193,131 @@ Views / Rules
Indexes
=======
* -Order duplicate index entries on creation by tid for faster heap lookups
* -Order duplicate index entries on creation by ctid for faster heap lookups
* Allow inherited tables to inherit index, UNIQUE constraint, and primary
key, foreign key [inheritance]
* UNIQUE INDEX on base column not honored on inserts from inherited table
INSERT INTO inherit_table (unique_index_col) VALUES (dup) should fail
[inheritance]
* UNIQUE INDEX on base column not honored on inserts/updates from
inherited table: INSERT INTO inherit_table (unique_index_col) VALUES
(dup) should fail [inheritance]
The main difficulty with this item is the problem of creating an index
that can spam more than one table.
* Add UNIQUE capability to non-btree indexes
* Add rtree index support for line, lseg, path, point
* Use indexes for min() and max() or convert to SELECT col FROM tab ORDER
BY col DESC LIMIT 1 if appropriate index exists and WHERE clause acceptible
* Use indexes for MIN() and MAX()
MIN/MAX queries can already be rewritten as SELECT col FROM tab ORDER
BY col {DESC} LIMIT 1. Completing this item involves making this
transformation automatically.
* Use index to restrict rows returned by multi-key index when used with
non-consecutive keys or OR clauses, so fewer heap accesses
* Be smarter about insertion of already-ordered data into btree index
non-consecutive keys to reduce heap accesses
For an index on col1,col2,col3, and a WHERE clause of col1 = 5 and
col3 = 9, spin though the index checking for col1 and col3 matches,
rather than just col1
* -Be smarter about insertion of already-ordered data into btree index
* Prevent index uniqueness checks when UPDATE does not modify the column
* Use bitmaps to fetch heap pages in sequential order [performance]
Uniqueness (index) checks are done when updating a column even if the
column is not modified by the UPDATE.
* Fetch heap pages matching index entries in sequential order [performance]
Rather than randomly accessing heap pages based on index entries, mark
heap pages needing access in a bitmap and do the lookups in sequential
order. Another method would be to sort heap ctids matching the index
before accessing the heap rows.
* Use bitmaps to combine existing indexes [performance]
Bitmap indexes allow single indexed columns to be combined to
dynamically create a composite index to match a specific query. Each
index is a bitmap, and the bitmaps are AND'ed or OR'ed to be combined.
* Allow use of indexes to search for NULLs
One solution is to create a partial index on an IS NULL expression.
* -Allow SELECT * FROM tab WHERE int2col = 4 to use int2col index, int8,
float4, numeric/decimal too
* Add FILLFACTOR to btree index creation
* Add concurrency to GIST
* Allow a single index to index multiple tables (for inheritance and subtables)
* Pack hash index buckets onto disk pages more efficiently
Currently no only one hash bucket can be stored on a page. Ideally
several hash buckets could be stored on a single page and greater
granularity used for the hash algorithm.
Commands
========
* Add BETWEEN ASYMMETRIC/SYMMETRIC (Christopher)
* Add BETWEEN ASYMMETRIC/SYMMETRIC
* Change LIMIT/OFFSET to use int8
* CREATE TABLE AS can not determine column lengths from expressions [atttypmod]
* Allow UPDATE to handle complex aggregates [update]
* Allow command blocks to ignore certain types of errors
* Allow UPDATE to handle complex aggregates [update] (?)
* -Allow command blocks to ignore certain types of errors
* Allow backslash handling in quoted strings to be disabled for portability
* Allow UPDATE, DELETE to handle table aliases for self-joins [delete]
The use of C-style backslashes (.e.g. \n, \r) in quoted strings is not
SQL-spec compliant, so allow such handling to be disabled.
* Allow DELETE to handle table aliases for self-joins [delete]
There is no way to specify use a table alias for the deleted table in
the DELETE WHERE clause because there is no FROM clause. Various
syntax extensions to add a FROM clause have been discussed. UPDATE
already has such an optional FROM clause.
* Add CORRESPONDING BY to UNION/INTERSECT/EXCEPT
* Allow REINDEX to rebuild all indexes, remove /contrib/reindex
* Allow REINDEX to rebuild all database indexes, remove /contrib/reindex
* Add ROLLUP, CUBE, GROUPING SETS options to GROUP BY
* Add schema option to createlang
* Allow savepoints / nested transactions [transactions] (Alvaro)
* Use nested transactions to prevent syntax errors from aborting a transaction
* Add a schema option to createlang
* -Allow savepoints / nested transactions [transactions] (Alvaro)
* -Use nested transactions to prevent syntax errors from aborting a transaction
* Allow UPDATE tab SET ROW (col, ...) = (...) for updating multiple columns
* Allow SET CONSTRAINTS to be qualified by schema/table
* Prevent COMMENT ON DATABASE from using a database name
* Allow SET CONSTRAINTS to be qualified by schema/table name
* -Prevent COMMENT ON DATABASE from using a database name
* -Add NO WAIT LOCKs
* Allow TRUNCATE ... CASCADE/RESTRICT
* Allow PREPARE of cursors
* Allow LISTEN/NOTIFY to store info in memory rather than tables
* Allow PREPARE to automatically determine parameter types based on the SQL
statement
* Allow LISTEN/NOTIFY to store info in memory rather than tables?
Currently LISTEN/NOTIFY information is stored in pg_listener. Storing
such information in memory would improve performance.
* -COMMENT ON [ CAST | CONVERSION | OPERATOR CLASS | LARGE OBJECT | LANGUAGE ]
(Christopher)
* Dump large object comments in custom dump format
* Add optional textual message to NOTIFY
This would allow an informational message to be added to the notify
message, perhaps indicating the row modified or other custom
information.
* -Allow more ISOLATION LEVELS to be accepted
* Allow CREATE TABLE foo (f1 INT CHECK (f1 > 0) CHECK (f1 < 10)) to work
by searching for non-conflicting constraint names, and prefix with table name
* Use more reliable method for CREATE DATABASE to get a consistent copy of db
by searching for non-conflicting constraint names, and prefix with
table name?
* Use more reliable method for CREATE DATABASE to get a consistent copy
of db?
Currently the system uses the operating system COPY command to create
new database.
* Add C code to copy directories for use in creating new databases
* -Have psql \dn show only visible temp schemas using current_schemas()
* -Have psql '\i ~/<tab><tab>' actually load files it displays from home dir
* Ignore temporary tables from other session when processing inheritance
* Ignore temporary tables from other sessions when processing
inheritance?
* -Add GUC setting to make created tables default to WITHOUT OIDS
* Have pg_ctl look at PGHOST in case it is a socket directory
* Allow column-level privileges
* Add a session mode to warn about non-standard SQL usage
* Have pg_ctl look at PGHOST in case it is a socket directory?
* Allow column-level GRANT/REVOKE privileges
* Add a session mode to warn about non-standard SQL usage in queries
* Add MERGE command that does UPDATE/DELETE, or on failure, INSERT (rules, triggers?)
* Add ON COMMIT capability to CREATE TABLE AS SELECT
* Add NOVICE output level for helpful messages like automatic sequence/index creation
@ -209,61 +328,106 @@ Commands
rows with DEFAULT value
o -ALTER TABLE ADD COLUMN column SERIAL doesn't create sequence because
of the item above
o Have ALTER TABLE rename SERIAL sequences
o Have ALTER TABLE RENAME rename SERIAL sequence names
o -Allow ALTER TABLE to modify column lengths and change to binary
compatible types
o Add ALTER DATABASE ... OWNER TO newowner
o -Add ALTER DATABASE ... OWNER TO newowner
o Add ALTER DOMAIN TYPE
o Allow ALTER TABLE ... ALTER CONSTRAINT ... RENAME
o Allow ALTER TABLE to change constraint deferrability and actions
o Disallow dropping of an inherited constraint
o Allow the schema of objects to be changed
o Add ALTER TABLESPACE to change location, name, owner
o Allow objects to be moved between tablespaces
o Allow objects to be moved to different schemas
o Allow ALTER TABLESPACE to move to different directories
o Allow databases, schemas, and indexes to be moved to different
tablespaces
o Allow moving system tables to other tablespaces, where possible
Currently non-global system tables must be in the default database
schema. Global system tables can never be moved.
o -Add ALTER DOMAIN, AGGREGATE, CONVERSION ... OWNER TO
o -Add ALTER SEQUENCE ... OWNER TO
* CLUSTER
o Automatically maintain clustering on a table
o Add ALTER TABLE table SET WITHOUT CLUSTER (Christopher)
This would require some background daemon to restore clustering
during periods of low usage. It might also require tables to be only
paritally filled for easier reorganization.
o -Add ALTER TABLE table SET WITHOUT CLUSTER (Christopher)
o Add default clustering to system tables
To do this, determine the ideal cluster index for each system
table and set the cluster setting during initdb.
* COPY
o -Allow dump/load of CSV format
o Allow COPY to report error lines and continue; optionally
allow error codes to be specified; requires savepoints or can
not be run in a multi-statement transaction
o Allow COPY to understand \x as hex
o Have COPY return number of rows loaded/unloaded
o Allow COPY to report error lines and continue
This requires the use of a savepoint before each COPY line is
processed, with ROLLBACK on COPY failure.
o Allow COPY to understand \x as a hex byte
o Have COPY return the number of rows loaded/unloaded (?)
* CURSOR
o Allow UPDATE/DELETE WHERE CURRENT OF cursor using per-cursor tid
stored in the backend (Gavin)
o Prevent DROP of table being referenced by our own open cursor
o Allow UPDATE/DELETE WHERE CURRENT OF cursor
This requires using the row ctid to map cursor rows back to the
original heap row. This become more complicated if WITH HOLD cursors
are to be supported because WITH HOLD cursors have a copy of the row
and no FOR UPDATE lock.
o Prevent DROP TABLE from dropping a row referenced by its own open
cursor (?)
o Allow pooled connections to list all open WITH HOLD cursors
Because WITH HOLD cursors exist outside transactions, this allows
them to be listed so they can be closed.
* INSERT
o Allow INSERT/UPDATE of system-generated oid value for a row
o Allow INSERT/UPDATE of the system-generated oid value for a row
o Allow INSERT INTO tab (col1, ..) VALUES (val1, ..), (val2, ..)
o Allow INSERT/UPDATE ... RETURNING new.col or old.col; handle
RULE cases (Philip)
o Allow INSERT/UPDATE ... RETURNING new.col or old.col
This is useful for returning the auto-generated key for an INSERT.
One complication is how to handle rules that run as part of
the insert.
* SHOW/SET
o Add SET PERFORMANCE_TIPS option to suggest INDEX, VACUUM, VACUUM
ANALYZE, and CLUSTER
o Add SET PATH for schemas
o Enforce rules for setting combinations
o Add SET PATH for schemas (?)
This is basically the same as SET search_path.
o Prevent conflicting SET options from being set
This requires a checking function to be called after the server
configuration file is read.
* SERVER-SIDE LANGUAGES
o Allow PL/PgSQL's RAISE function to take expressions
o Allow PL/PgSQL's RAISE function to take expressions (?)
Currently only constants are supported.
o Change PL/PgSQL to use palloc() instead of malloc()
o -Allow Java server-side programming
o Fix problems with complex temporary table creation/destruction
without using PL/PgSQL EXECUTE, needs cache prevention/invalidation
o Handle references to temporary tables that are created, destroyed,
then recreated during a session, and EXECUTE is not used
This requires the cached PL/PgSQL byte code to be invalidated when
an object referenced in the function is changed.
o Fix PL/pgSQL RENAME to work on variables other than OLD/NEW
o Improve PL/PgSQL exception handling
o Improve PL/PgSQL exception handling using savepoints
o -Allow PL/pgSQL parameters to be specified by name and type during definition
o Allow function parameters to be passed by name,
get_employee_salary(emp_id => 12345, tax_year => 2001)
o Add PL/PgSQL packages
o Add table function support to pltcl, plperl, plpython
o Add Oracle-style packages
o Add table function support to pltcl, plperl, plpython (?)
o Allow PL/pgSQL to name columns by ordinal position, e.g. rec.(3)
o Allow PL/pgSQL EXECUTE query_var INTO record_var;
o Add capability to create and call PROCEDURES
@ -273,28 +437,42 @@ Commands
Clients
=======
* Add XML capability to pg_dump and COPY, when backend XML capability
* Add XML output to pg_dump and COPY
We already allow XML to be stored in the database, and XPath queries
can be used on that data using /contrib/xml2. It also supports XSLT
transformations.
* -Allow psql \du to show users, and add \dg for groups
* Allow clients to query a list of WITH HOLD cursors and prepared statements
* Add a libpq function to support Parse/DescribeStatement capability
* Prevent libpq's PQfnumber() from lowercasing the column name
* Prevent libpq's PQfnumber() from lowercasing the column name (?)
* -Allow pg_dump to dump CREATE CONVERSION (Christopher)
* Allow libpq to return information about prepared queries
* -Make pg_restore continue after errors, so it acts more like pg_dump scripts
* Have psql show current values for a sequence
* Allow pg_dumpall to use non-text output formats
* Have pg_dump use multi-statement transactions for INSERT dumps
* Move psql backslash database information into the backend, use mnemonic
commands? [psql]
* Allow pg_dump to use multiple -t and -n switches
* Fix oid2name and dbsize for tablespaces
* Consistenly display privilege information for all objects in psql
* ECPG
This would allow non-psql clients to pull the same information out of
the database as psql.
* Allow pg_dump to use multiple -t and -n switches
This should be done by allowing a '-t schema.table' syntax.
* Fix oid2name and dbsize for tablespaces
* Consistently display privilege information for all objects in psql
* ECPG (?)
o Docs
o Implement set descriptor, using descriptor
o Solve cardinality > 1 for input descriptors / variables
o Improve error handling
Document differences between ecpg and the SQL standard and
information about the Informix-compatibility module.
o -Implement SET DESCRIPTOR
o Solve cardinality > 1 for input descriptors / variables (?)
o Improve error handling (?)
o Add a semantic check level, e.g. check if a table really exists
o fix handling of DB attributes that are arrays
o Use backend PREPARE/EXECUTE facility for ecpg where possible
@ -305,34 +483,55 @@ Clients
o Allow multidimensional arrays
Referential Integrity
=====================
* Add MATCH PARTIAL referential integrity
* Add deferred trigger queue file (Jan)
* Implement dirty reads or shared row locks and use them in RI triggers
* Add deferred trigger queue file
Right now all deferred trigger information is stored in backend
memory. This could exhaust memory for very large trigger queues.
This item involves dumping large queues into files.
* Implement dirty reads or shared row locks and use them in RI triggers (?)
* Enforce referential integrity for system tables
* Change foreign key constraint for array -> element to mean element
in array
* Allow DEFERRABLE UNIQUE constraints
in array (?)
* Allow DEFERRABLE UNIQUE constraints (?)
* Allow triggers to be disabled [trigger]
Currently the only way to disable triggers is to modify the system
tables.
* With disabled triggers, allow pg_dump to use ALTER TABLE ADD FOREIGN KEY
If the dump is known to be valid, allow foreign keys to be added
without revalidating the data.
* Allow statement-level triggers to access modified rows
* Support triggers on columns (Neil)
* Support triggers on columns
* Have AFTER triggers execute after the appropriate SQL statement in a
function, not at the end of the function
* -Print table names with constraint names in error messages, or make constraint
names unique within a schema
* -Issue NOTICE if foreign key data requires costly test to match primary key
* Remove CREATE CONSTRAINT TRIGGER
This was used in older releases to dump referential integrity
constraints.
* Allow AFTER triggers on system tables
System tables are modified in many places in the backend without going
through the executor and therefore not causing triggers to fire. To
complete this item, the functions that modify system tables will have
to fire triggers.
Dependency Checking
===================
* Flush cached query plans when their underlying catalog data changes
* Flush cached query plans when the dependent objects change
* -Use dependency information to dump data in proper order
* -Have pg_dump -c clear the database using dependency information
@ -340,15 +539,29 @@ Dependency Checking
Exotic Features
===============
* Add SQL99 WITH clause to SELECT (Tom, Fernando)
* Add SQL99 WITH RECURSIVE to SELECT (Tom, Fernando)
* Add pre-parsing phase that converts non-ANSI features to supported features
* Add SQL99 WITH clause to SELECT
* Add SQL99 WITH RECURSIVE to SELECT
* Add pre-parsing phase that converts non-ANSI syntax to supported
syntax
This could allow SQL written for other databases to run without
modification.
* Allow plug-in modules to emulate features from other databases
* SQL*Net listener that makes PostgreSQL appear as an Oracle database
to clients
* Add two-phase commit to all distributed transactions with
offline/readonly server status or administrator notification for failure
* Allow cross-db queries with transaction semantics
* Allow queries across databases or servers with transaction
semantics
Right now contrib/dblink can be used to issue such queries except it
does not have locking or transaction semantics. Two-phase commit is
needed to enable transaction semantics.
* Add two-phase commit
This will involve adding a way to respond to commit failure by either
taking the server into offline/readonly mode or notifying the
administrator
PERFORMANCE
@ -358,45 +571,80 @@ PERFORMANCE
Fsync
=====
* Delay fsync() when other backends are about to commit too
o Determine optimal commit_delay value
* Improve commit_delay handling to reduce fsync()
* Determine optimal fdatasync/fsync, O_SYNC/O_DSYNC options
o Allow multiple blocks to be written to WAL with one write()
* Allow multiple blocks to be written to WAL with one write()
* Add an option to sync() before fsync()'ing checkpoint files
Cache
=====
* Shared catalog cache, reduce lseek()'s by caching table size in shared area
* Add free-behind capability for large sequential scans [fadvise]
* Consider use of open/fcntl(O_DIRECT) to minimize OS caching
* Cache last known per-tuple offsets to speed long tuple access, adjusting
for NULLs and TOAST values
* Use a fixed row count and a +/- count with MVCC visibility rules
to allow fast COUNT(*) queries with no WHERE clause(?) [count]
* Cache last known per-tuple offsets to speed long tuple access
While column offsets are already cached, the cache can not be used if
the tuple has NULLs or TOAST columns because these values change the
typical column offsets. Caching of such offsets could be accomplished
by remembering the previous offsets and use them again if the row has
the same pattern.
* Speed up COUNT(*)
We could use a fixed row count and a +/- count to follow MVCC
visibility rules, or a single cached value could be used and
invalidated if anyone modifies the table. [count]
Vacuum
======
* Improve speed with indexes (perhaps recreate index instead)
* Improve speed with indexes
For large table adjustements during vacuum, it is faster to reindex
rather than update the index.
* Reduce lock time by moving tuples with read lock, then write
lock and truncate table
* Provide automatic running of vacuum in the background in backend
Moved tuples are invisible to other backends so they don't require a
write lock. However, the read lock promotion to write lock could lead
to deadlock situations.
* -Provide automatic running of vacuum in the background in backend
rather than in /contrib (Matthew)
* Allow free space map to be auto-sized or warn when it is too small
* Maintain a map of recently-expired of pages so vacuum can reclaim
free space without a sequential scan
* Have VACUUM FULL use REINDEX rather than index vacuum
The free space map is in shared memory so resizing is difficult.
* Maintain a map of recently-expired rows
This allows vacuum to reclaim free space without requiring
a sequential scan
Locking
=======
* Make locking of shared data structures more fine-grained
This requires that more locks be acquired but this would reduce lock
contention, improving concurrency.
* Add code to detect an SMP machine and handle spinlocks accordingly
from distributted.net, http://www1.distributed.net/source,
in client/common/cpucheck.cpp
On SMP machines, it is possible that locks might be released shortly,
while on non-SMP machines, the backend should sleep so the process
holding the lock can complete and release it.
* Improve SMP performance on i386 machines
i386-based SMP machines can generate excessive context switching
caused by lock failure in high concurrency situations. This may be
caused by CPU cache line invalidation inefficiencies.
* Research use of sched_yield() for spinlock acquisition failure
@ -404,60 +652,91 @@ Startup Time
============
* Experiment with multi-threaded backend [thread]
This would prevent the overhead associated with process creation. Most
operating systems have trivial process creation time compared to
database startup overhead, but a few operating systems (WIn32,
Solaris) might benefit from threading.
* Add connection pooling [pool]
* Allow persistent backends [pool]
* Create a transaction processor to aid in persistent connections and
connection pooling [pool]
* Do listen() in postmaster and accept() in pre-forked backend
* Have pre-forked backend pre-connect to last requested database or pass
file descriptor to backend pre-forked for matching database
It is unclear if this should be done inside the backend code or done
by something external like pgpool. The passing of file descriptors to
existing backends is one of the difficulties with a backend approach.
Write-Ahead Log
===============
* Have after-change WAL write()'s write only modified data to kernel
* Reduce number of after-change WAL writes; they exist only to gaurd against
partial page writes [wal]
* Turn off after-change writes if fsync is disabled (?)
* Eliminate need to write full pages to WAL before page modification [wal]
Currently, to protect against partial disk page writes, we write the
full page images to WAL before they are modified so we can correct any
partial page writes during recovery.
* Reduce WAL traffic so only modified values are written rather than
entire rows (?)
* Turn off after-change writes if fsync is disabled
If fsync is off, there is no purpose in writing full pages to WAL
* Add WAL index reliability improvement to non-btree indexes
* Find proper defaults for postgresql.conf WAL entries
* Allow xlog directory location to be specified during initdb, perhaps
using symlinks
* Allow the pg_xlog directory location to be specified during initdb
with a symlink back to the /data location
* Allow WAL information to recover corrupted pg_controldata
* Find a way to reduce rotational delay when repeatedly writing
last WAL page
Currently fsync of WAL requires the disk platter to perform a full
rotation to fsync again. One idea is to write the WAL to different
offsets that might reduce the rotational delay.
Optimizer / Executor
====================
* Missing optimizer selectivities for date, r-tree, etc
* Allow ORDER BY ... LIMIT to select top values without sort or index
using a sequential scan for highest/lowest values (Oleg)
* Precompile SQL functions to avoid overhead (Neil)
* Add missing optimizer selectivities for date, r-tree, etc
* Allow ORDER BY ... LIMIT 1 to select high/low value without sort or
index using a sequential scan for highest/lowest values
If only one value is needed, there is no need to sort the entire
table. Instead a sequential scan could get the matching value.
* Precompile SQL functions to avoid overhead
* Add utility to compute accurate random_page_cost value
* Improve ability to display optimizer analysis using OPTIMIZER_DEBUG
* Use CHECK constraints to improve optimizer decisions
* Check GUC geqo_threshold to see if it is still accurate
* Allow sorting, temp files, temp tables to use multiple work directories
* Improve the planner to use CHECK constraints to prune the plan (for subtables)
This allows the I/O load to be spread across multiple disk drives.
* Have EXPLAIN ANALYZE highlight poor optimizer estimates
* Use CHECK constraints to influence optimizer decisions
CHECK constraints contain information about the distribution of values
within the table. This is also useful for implementing subtables where
a tables content is distributed across several subtables.
Miscellaneous
=============
* Do async I/O for faster random read-ahead of data
Async I/O allows multiple I/O requests to be sent to the disk with
results coming back asynchronously.
* Use mmap() rather than SYSV shared memory or to write WAL files (?) [mmap]
* Improve caching of attribute offsets when NULLs exist in the row
This would remove the requirement for SYSV SHM but would introduce
portability issues. Anonymous mmap is required to prevent I/O
overhead.
* Add a script to ask system configuration questions and tune postgresql.conf
* Allow partitioning of table into multiple subtables
* -Use background process to write dirty shared buffers to disk
* Investigate SMP context switching issues
* Use a phantom command counter for nested subtransactions to reduce
tuple overhead
Source Code
===========
@ -467,69 +746,63 @@ Source Code
* Remove warnings created by -Wcast-align
* Move platform-specific ps status display info from ps_status.c to ports
* Improve access-permissions check on data directory in Cygwin (Tom)
* Add documentation for perl, including mention of DBI/DBD perl location
* Create improved PostgreSQL introductory documentation for the PHP
manuals
* Add optional CRC checksum to heap and index pages
* -Change representation of whole-tuple parameters to functions
* Clarify use of 'application' and 'command' tags in SGML docs
* Better document ability to build only certain interfaces (Marc)
* Remove or relicense modules that are not under the BSD license, if possible
* Remove memory/file descriptor freeing before ereport(ERROR) (Bruce)
* Remove memory/file descriptor freeing before ereport(ERROR)
* Acquire lock on a relation before building a relcache entry for it
* Research interaction of setitimer() and sleep() used by statement_timeout
* -Add checks for fclose() failure (Tom)
* -Change CVS ID to PostgreSQL
* -Exit postmaster if postgresql.conf can not be opened
* Rename /scripts directory because they are all C programs now
* Allow creation of a libpq-only tarball
* Promote debug_query_string into a server-side function current_query()
* Allow the identifier length to be increased via a configure option
* Improve CREATE SCHEMA regression test
* Allow binaries to be statically linked so they are more easily relocated
* Wire Protocol Changes
o Dynamic character set handling
o Allow dynamic character set handling
o Add decoded type, length, precision
o Compression?
o Use compression?
o Update clients to use data types, typmod, schema.table.column names of
result sets using new query protocol
---------------------------------------------------------------------------
Developers who have claimed items are:
--------------------------------------
* Alvaro is Alvaro Herrera <alvherre@dcc.uchile.cl>
* Barry is Barry Lind <barry@xythos.com>
* Billy is Billy G. Allie <Bill.Allie@mug.org>
* Andrew is Andrew Dunstan
* Bruce is Bruce Momjian <pgman@candle.pha.pa.us> of Software Research Assoc.
* Christopher is Christopher Kings-Lynne <chriskl@familyhealth.com.au> of
Family Health Network
* Claudio is ?
* D'Arcy is D'Arcy J.M. Cain <darcy@druid.net> of The Cain Gang Ltd.
* Dave is Dave Cramer <dave@fastcrypt.com>
* Edmund is Edmund Mergl <E.Mergl@bawue.de>
* Fernando is Fernando Nasser <fnasser@redhat.com> of Red Hat
* Fabien is Fabien Coelho
* Gavin is Gavin Sherry <swm@linuxworld.com.au> of Alcove Systems Engineering
* Greg is Greg Sabino Mullane <greg@turnstep.com>
* Hiroshi is Hiroshi Inoue <Inoue@tpf.co.jp>
* Karel is Karel Zak <zakkr@zf.jcu.cz>
* Jan is Jan Wieck <JanWieck@Yahoo.com> of Afilias, Inc.
* Joe is Joe Conway <mail@joeconway.com>
* Liam is Liam Stewart <liams@redhat.com> of Red Hat
* Karel is Karel Zak <zakkr@zf.jcu.cz>
* Kris is Kris Jurka
* Magnus is Magnus Haglander (?)
* Manfred is Manfred Koizar <
* Marc is Marc Fournier <scrappy@hub.org> of PostgreSQL, Inc.
* Mark is Mark Hollomon <mhh@mindspring.com>
* Matthew T. O'Connor <matthew@zeut.net>
* Michael is Michael Meskes <meskes@postgresql.org> of Credativ
* Neil is Neil Conway <neilc@samurai.com>
* Oleg is Oleg Bartunov <oleg@sai.msu.su>
* Peter M is Peter T Mount <peter@retep.org.uk> of Retep Software
* Peter E is Peter Eisentraut <peter_e@gmx.net>
* Peter is Peter Eisentraut <peter_e@gmx.net>
* Philip is Philip Warner <pjw@rhyme.com.au> of Albatross Consulting Pty. Ltd.
* Rod is Rod Taylor <pg@rbt.ca>
* Ross is Ross J. Reedstrom <reedstrm@wallace.ece.rice.edu>
* Simon is Simon Riggs
* Stephan is Stephan Szabo <sszabo@megazone23.bigpanda.com>
* Tatsuo is Tatsuo Ishii <t-ishii@sra.co.jp> of Software Research Assoc.
* Thomas is Thomas Lockhart <lockhart@fourpalms.org> of Jet Propulsion Labratory
* Teodor is
* Tom is Tom Lane <tgl@sss.pgh.pa.us> of Red Hat