gem-hunter
Language: C
Copy a Postgres database to a target Postgres server (pg_dump | pg_restore on steroids)
pgcopydb is a tool that automates running pg_dump | pg_restore between two
running Postgres servers. To make a copy of a database to another server as
quickly as possible, one would like to use the parallel options of pg_dump
and still be able to stream the data to as many pg_restore jobs.
The idea would be to use pg_dump --jobs=N --format=directory postgres://user@source/dbname | pg_restore --jobs=N --format=directory -d postgres://user@target/dbname in a way. This command line can't be made to
work, unfortunately, because pg_dump --format=directory writes to local
files and directories first, and then later pg_restore --format=directory
can be used to read from those files again.
Given that, pgcopydb then uses pg_dump and pg_restore for the schema parts of the process, and implements its own data copying multi-process streaming parts. Also, pgcopydb bypasses pg_restore index building and drives that internally so that all indexes may be built concurrently.
pgcopydb implements both the base copy of a database and also Change Data
Capture to allow replay of changes from the source database to the target
database. The Change Data Capture facility is implemented using Postgres
Logical Decoding infrastructure. The default plugin is pgoutput, which
is built into PostgreSQL core (since version 10) and requires no extension
installation. The test_decoding and wal2json plugins are also
supported.
WAL changes are stored in structured SQLite databases (output.db and
replay.db) in the pgcopydb work directory, enabling reliable processing
across process restarts, size-based file rotation, and a stream prune
command to reclaim disk space.
The pgcopydb follow command implements a logical replication client for
the pgoutput logical decoding plugin (built into PostgreSQL core), as
well as test_decoding and wal2json.
The pgcopydb clone --follow command implements a full solution for online
migration. Beware that online migrations involve a lot more complexities
when compared to offline migration. It is always a good idea to first
implement offline migration first. The command pgcopydb clone is used to
implement the offline migration approach.
The pgcopydb clone --all-databases command copies an entire Postgres
instance (all databases) from source to target in a single operation.
Full documentation is available online, including manual pages of all the pgcopydb sub-commands. Check out https://pgcopydb.readthedocs.io/.
$ pgcopydb help
pgcopydb
clone Clone an entire database from source to target
fork Clone an entire database from source to target
follow Replay changes from the source database to the target database
snapshot Create and export a snapshot on the source database
+ compare Compare source and target databases
+ copy Implement the data section of the database copy
+ dump Dump database objects from a Postgres instance
+ restore Restore database objects into a Postgres instance
+ list List database objects from a Postgres instance
+ stream Stream changes from the source database
ping Attempt to connect to the source and target instances
help Print help message
version Print pgcopydb version
pgcopydb compare
schema Compare source and target schema
data Compare source and target data
pgcopydb copy
db Copy an entire database from source to target
roles Copy the roles from the source instance to the target instance
extensions Copy the extensions from the source instance to the target instance
schema Copy the database schema from source to target
data Copy the data section from source to target
table-data Copy the data from all tables in database from source to target
blobs Copy the blob data from the source database to the target
sequences Copy the current value from all sequences in database from source to target
indexes Create all the indexes found in the source database in the target
constraints Create all the constraints found in the source database in the target
pgcopydb dump
schema Dump source database schema as custom files in work directory
roles Dump source database roles as custom file in work directory
pgcopydb restore
schema Restore a database schema from custom files to target database
pre-data Restore a database pre-data schema from custom file to target database
post-data Restore a database post-data schema from custom file to target database
roles Restore database roles from SQL file to target database
pgcopydb list
databases List databases
extensions List all the source extensions to copy
collations List all the source collations to copy
tables List all the source tables to copy data from
table-parts List a source table copy partitions
sequences List all the source sequences to copy data from
indexes List all the indexes to create again after copying the data
depends List all the dependencies to filter-out
schema List the schema to migrate, formatted in JSON
progress List the progress
toc List and filter a pg_dump archive's Table of Contents
pgcopydb stream
setup Setup source and target systems for logical decoding
cleanup Cleanup source and target systems for logical decoding
prune Remove already-applied CDC files from disk to reclaim disk space
event-trigger Install the event trigger that emits DDL as logical decoding messages
prefetch Stream changes from the source database into the SQLite CDC store
catchup Transform and apply prefetched changes from the SQLite CDC store to the target
replay Replay changes from the source to the target database, live
+ sentinel Maintain a sentinel table
receive Stream changes from the source database
apply Apply changes from the replayDB to the target database, or stdout
pgcopydb stream sentinel
setup Setup the sentinel table
get Get the sentinel table values
+ set Set the sentinel table values
pgcopydb stream sentinel set
startpos Set the sentinel start position LSN
endpos Set the sentinel end position LSN
apply Set the sentinel apply mode
prefetch Set the sentinel prefetch mode
When using pgcopydb it is possible to achieve the result outlined before
with this simple command line:
$ export PGCOPYDB_SOURCE_PGURI="postgres://user@source.host.dev/dbname" $ export PGCOPYDB_TARGET_PGURI="postgres://role@target.host.dev/dbname" $ pgcopydb clone --table-jobs 8 --index-jobs 2
A typical output from the command would contain lots of lines of logs, and then a table summary with a line per table and some information (timing for the table COPY, cumulative timing for the CREATE INDEX commands), and then an overall summary that looks like the following:
22:37:01.564 22 INFO Running pgcopydb version 0.18 from "/usr/bin/pgcopydb"
22:37:01.565 22 INFO [SOURCE] Copying database from "postgres://pagila@source/pagila?keepalives=1&keepalives_idle=10&keepalives_interval=10&keepalives_count=60"
22:37:01.565 22 INFO [TARGET] Copying database into "postgres://pagila@target/pagila?keepalives=1&keepalives_idle=10&keepalives_interval=10&keepalives_count=60"
22:37:01.582 22 INFO Exported snapshot "00000003-0000002D-1" from the source database
22:37:01.583 24 INFO STEP 1: fetch source database tables, indexes, and sequences
22:37:01.737 24 INFO Fetched information for 21 tables (including 11 tables split in 30 partitions total), with an estimated total of 49 636 tuples and 5224 kB on-disk
22:37:01.743 24 INFO Fetched information for 54 indexes (supporting 21 constraints)
22:37:01.756 24 INFO Fetched information for 7 extensions
22:37:01.795 24 INFO STEP 2: dump the source database schema (pre/post data)
22:37:01.796 24 INFO /usr/bin/pg_dump -Fc --snapshot 00000003-0000002D-1 --section=pre-data --section=post-data --file /tmp/pgcopydb/schema/schema.dump 'postgres://pagila@source/pagila?keepalives=1&keepalives_idle=10&keepalives_interval=10&keepalives_count=60'
22:37:01.846 24 INFO STEP 3: restore the pre-data section to the target database
22:37:01.862 24 INFO /usr/bin/pg_restore --dbname 'postgres://pagila@target/pagila?keepalives=1&keepalives_idle=10&keepalives_interval=10&keepalives_count=60' --section pre-data --jobs 4 --use-list /tmp/pgcopydb/schema/pre-filtered.list /tmp/pgcopydb/schema/schema.dump
22:37:01.956 34 INFO STEP 4: starting 4 table-data COPY processes
22:37:01.956 36 INFO STEP 8: starting 4 VACUUM processes
22:37:01.959 24 INFO STEP 9: reset sequences values
22:37:01.965 35 INFO STEP 6: starting 4 CREATE INDEX processes
22:37:01.965 35 INFO STEP 7: constraints are built by the CREATE INDEX processes
22:37:02.895 24 INFO All step are now done, 1s100 elapsed
22:37:02.895 24 INFO Printing summary for 21 tables and 54 indexes
OID | Schema | Name | Parts | copy duration | transmitted bytes | indexes | create index duration
------+--------+------------------+-------+---------------+-------------------+---------+----------------------
17182 | public | rental | 5 | 250ms | 5477 kB | 3 | 34ms
17071 | public | film | 4 | 124ms | 1336 kB | 5 | 26ms
17130 | public | inventory | 4 | 104ms | 680 kB | 2 | 17ms
17082 | public | film_actor | 2 | 60ms | 320 kB | 2 | 15ms
17200 | public | staff | 3 | 96ms | 718 kB | 1 | 5ms
17157 | public | payment_p2022_02 | 2 | 46ms | 289 kB | 4 | 28ms
17161 | public | payment_p2022_03 | 2 | 60ms | 326 kB | 4 | 20ms
17165 | public | payment_p2022_04 | 2 | 28ms | 306 kB | 4 | 22ms
17169 | public | payment_p2022_05 | 2 | 54ms | 322 kB | 4 | 26ms
17173 | public | payment_p2022_06 | 2 | 40ms | 319 kB | 4 | 25ms
17177 | public | payment_p2022_07 | 2 | 60ms | 280 kB | 1 | 6ms
17086 | public | film_category | 1 | 14ms | 69 kB | 1 | 6ms
17043 | public | customer | 1 | 15ms | 67 kB | 4 | 20ms
17096 | public | address | 1 | 17ms | 55 kB | 2 | 7ms
17153 | public | payment_p2022_01 | 1 | 16ms | 43 kB | 4 | 19ms
17104 | public | city | 1 | 13ms | 24 kB | 2 | 9ms
17209 | public | store | 1 | 10ms | 18 kB | 2 | 11ms
17055 | public | actor | 1 | 14ms | 8328 B | 2 | 13ms
17063 | public | category | 1 | 15ms | 540 B | 1 | 3ms
17112 | public | country | 1 | 14ms | 3829 B | 1 | 5ms
17136 | public | language | 1 | 14ms | 297 B | 1 | 6ms
Step Connection Duration Transfer Concurrency
-------------------------------------------------- ---------- ---------- ---------- ------------
Catalog Queries (table ordering, filtering, etc) source 124ms 1
Dump Schema source 50ms 1
Prepare Schema target 94ms 1
COPY, INDEX, CONSTRAINTS, VACUUM (wall clock) both 834ms 12
COPY (cumulative) both 455ms 3213 kB 4
CREATE INDEX (cumulative) target 318ms 4
CONSTRAINTS (cumulative) target 5ms 4
VACUUM (cumulative) target 185ms 4
Reset Sequences both 12ms 1
Large Objects (cumulative) both 0ms 4
Finalize Schema both 112ms 4
-------------------------------------------------- ---------- ---------- ---------- ------------
Total Wall Clock Duration both 1s100 20
See our documentation.
The reason why pgcopydb has been developed is mostly to allow two aspects
that are not possible to achieve directly with pg_dump and pg_restore,
and that requires just enough fiddling around that not many scripts have
been made available to automate around.
First aspect is that for pg_dump and pg_restore to implement concurrency
they need to write to an intermediate file first.
The docs for
pg_dump say the
following about the --jobs parameter:
You can only use this option with the directory output format because this is the only output format where multiple processes can write their data at the same time.
The docs for
pg_restore say
the following about the --jobs parameter:
Only the custom and directory archive formats are supported with this option. The input must be a regular file or directory (not, for example, a pipe or standard input).
So the first idea with pgcopydb is to provide the --jobs concurrency and
bypass intermediate files (and directories) altogether, at least as far as
the actual TABLE DATA set is concerned.
The trick to achieve that is that pgcopydb must be able to connect to the
source database during the whole operation, when pg_restore may be used
from an export on-disk, without having to still be able to connect to the
source database. In the context of pgcopydb requiring access to the source
database is fine. In the context of pg_restore, it would not be
acceptable.
The other aspect that pg_dump and pg_restore are not very smart about is
how they deal with the indexes that are used to support constraints, in
particular unique constraints and primary keys.
Those indexes are exported using the ALTER TABLE command directly. This is
fine because the command creates both the constraint and the underlying
index, so the schema in the end is found as expected.
That said, those ALTER TABLE ... ADD CONSTRAINT commands require a level
of locking that prevents any concurrency. As we can read on the docs for
ALTER TABLE:
Although most forms of ADD table_constraint require an ACCESS EXCLUSIVE lock, ADD FOREIGN KEY requires only a SHARE ROW EXCLUSIVE lock. Note that ADD FOREIGN KEY also acquires a SHARE ROW EXCLUSIVE lock on the referenced table, in addition to the lock on the table on which the constraint is declared.
The trick is then to first issue a CREATE UNIQUE INDEX statement and when
the index has been built then issue a second command in the form of ALTER TABLE ... ADD CONSTRAINT ... PRIMARY KEY USING INDEX ..., as in the
following example taken from the logs of actually running pgcopydb:
...
21:52:06 68898 INFO COPY "demo"."tracking";
21:52:06 68899 INFO COPY "demo"."client";
21:52:06 68899 INFO Creating 2 indexes for table "demo"."client"
21:52:06 68906 INFO CREATE UNIQUE INDEX client_pkey ON demo.client USING btree (client);
21:52:06 68907 INFO CREATE UNIQUE INDEX client_pid_key ON demo.client USING btree (pid);
21:52:06 68898 INFO Creating 1 indexes for table "demo"."tracking"
21:52:06 68908 INFO CREATE UNIQUE INDEX tracking_pkey ON demo.tracking USING btree (client, ts);
21:52:06 68907 INFO ALTER TABLE "demo"."client" ADD CONSTRAINT "client_pid_key" UNIQUE USING INDEX "client_pid_key";
21:52:06 68906 INFO ALTER TABLE "demo"."client" ADD CONSTRAINT "client_pkey" PRIMARY KEY USING INDEX "client_pkey";
21:52:06 68908 INFO ALTER TABLE "demo"."tracking" ADD CONSTRAINT "tracking_pkey" PRIMARY KEY USING INDEX "tracking_pkey";
...
This trick is worth a lot of performance gains on its own, as has been discovered and experienced and appreciated by pgloader users already.
At run-time pgcopydb depends on the pg_dump and pg_restore tools being
available in the PATH. The tools version should match the Postgres version
of the target database.
When you have multiple versions of Postgres installed, consider exporting
the PG_CONFIG environment variable to the version you want to use.
pgcopydb then uses the PG_CONFIG from the path and runs ${PG_CONFIG} --bindir to find the pg_dump and pg_restore binaries it needs.
The pgcopydb command line also includes entry points that allows
implementing any step on its own.
pgcopydb snapshot &pgcopydb dump schemapgcopydb restore pre-datapgcopydb copy table-datapgcopydb copy blobspgcopydb copy sequencespgcopydb copy indexespgcopydb copy constraintspgcopydb restore post-datakill %1Using individual commands fails to provide the advanced concurrency
capabilities of the main pgcopydb clone command, so it is strongly
advised to prefer that main command.
Also when using separate commands, one has to consider the --snapshot
option that allows for consistent operations. A background process should
then export the snapshot and maintain a transaction opened for the duration
of the operations. See documentation for pgcopydb snapshot.
Copyright (c) The PostgreSQL Global Development Group.
This project is licensed under the PostgreSQL License, see LICENSE file for details.
This project includes bundled third-party dependencies, see NOTICE file for details.
This project is maintained through OSS Maintenance — a funded model designed for enterprises running these tools in production.
PostgreSQL consultancies can also subscribe at the Partner tier to cover their entire client base under a single plan.