Postgres partitioning as easy as pie. Works great for both new and existing tables, with zero downtime and minimal app changes. No need to install anything on your database server. Archive older data on a rolling basis to keep your database size under control.
🍊 Battle-tested at Instacart
pgslice is a command line tool. To install, run:
gem install pgslice
This will give you the pgslice
command. You can also install it with Homebrew or Docker. If installation fails, you may need to install dependencies.
-
Ensure the table you want to partition has been created. We’ll refer to this as
<table>
. -
Specify your database credentials
export PGSLICE_URL=postgres:https://localhost/myapp_development
- Create an intermediate table
pgslice prep <table> <column> <period>
Period can be day
, month
, or year
.
This creates a partitioned table named <table>_intermediate
.
- Add partitions to the intermediate table
pgslice add_partitions <table> --intermediate --past 3 --future 3
Use the --past
and --future
options to control the number of partitions.
- Optional, for tables with data - Fill the partitions in batches with data from the original table
pgslice fill <table>
Use the --batch-size
and --sleep
options to control the speed.
To sync data across different databases, check out pgsync.
- Analyze tables
pgslice analyze <table>
- Swap the intermediate table with the original table
pgslice swap <table>
The original table is renamed <table>_retired
and the intermediate table is renamed <table>
.
- Fill the rest (rows inserted between the first fill and the swap)
pgslice fill <table> --swapped
- Back up the retired table with a tool like pg_dump and drop it
pg_dump -c -Fc -t <table>_retired $PGSLICE_URL > <table>_retired.dump
psql -c "DROP TABLE <table>_retired" $PGSLICE_URL
pgslice prints the SQL commands that were executed on the server. To print without executing, use the --dry-run
option.
pgslice prep visits created_at month
BEGIN;
CREATE TABLE "public"."visits_intermediate" (LIKE "public"."visits" INCLUDING DEFAULTS INCLUDING CONSTRAINTS INCLUDING STORAGE INCLUDING COMMENTS) PARTITION BY RANGE ("created_at");
CREATE INDEX ON "public"."visits_intermediate" USING btree ("created_at");
COMMENT ON TABLE "public"."visits_intermediate" is 'column:createdAt,period:day,cast:date,version:3';
COMMIT;
pgslice add_partitions visits --intermediate --past 1 --future 1
BEGIN;
CREATE TABLE "public"."visits_202108" PARTITION OF "public"."visits_intermediate" FOR VALUES FROM ('2021-08-01') TO ('2021-09-01');
ALTER TABLE "public"."visits_202108" ADD PRIMARY KEY ("id");
CREATE TABLE "public"."visits_202109" PARTITION OF "public"."visits_intermediate" FOR VALUES FROM ('2021-09-01') TO ('2021-10-01');
ALTER TABLE "public"."visits_202109" ADD PRIMARY KEY ("id");
CREATE TABLE "public"."visits_202110" PARTITION OF "public"."visits_intermediate" FOR VALUES FROM ('2021-10-01') TO ('2021-11-01');
ALTER TABLE "public"."visits_202110" ADD PRIMARY KEY ("id");
COMMIT;
pgslice fill visits
/* 1 of 3 */
INSERT INTO "public"."visits_intermediate" ("id", "user_id", "ip", "created_at")
SELECT "id", "user_id", "ip", "created_at" FROM "public"."visits"
WHERE "id" > 0 AND "id" <= 10000 AND "created_at" >= '2021-08-01'::date AND "created_at" < '2021-11-01'::date
/* 2 of 3 */
INSERT INTO "public"."visits_intermediate" ("id", "user_id", "ip", "created_at")
SELECT "id", "user_id", "ip", "created_at" FROM "public"."visits"
WHERE "id" > 10000 AND "id" <= 20000 AND "created_at" >= '2021-08-01'::date AND "created_at" < '2021-11-01'::date
/* 3 of 3 */
INSERT INTO "public"."visits_intermediate" ("id", "user_id", "ip", "created_at")
SELECT "id", "user_id", "ip", "created_at" FROM "public"."visits"
WHERE "id" > 20000 AND "id" <= 30000 AND "created_at" >= '2021-08-01'::date AND "created_at" < '2021-11-01'::date
pgslice analyze visits
ANALYZE VERBOSE "public"."visits_202108";
ANALYZE VERBOSE "public"."visits_202109";
ANALYZE VERBOSE "public"."visits_202110";
ANALYZE VERBOSE "public"."visits_intermediate";
pgslice swap visits
BEGIN;
SET LOCAL lock_timeout = '5s';
ALTER TABLE "public"."visits" RENAME TO "visits_retired";
ALTER TABLE "public"."visits_intermediate" RENAME TO "visits";
ALTER SEQUENCE "public"."visits_id_seq" OWNED BY "public"."visits"."id";
COMMIT;
To add partitions, use:
pgslice add_partitions <table> --future 3
Add this as a cron job to create a new partition each day, month, or year.
# day
0 0 * * * pgslice add_partitions <table> --future 3 --url ...
# month
0 0 1 * * pgslice add_partitions <table> --future 3 --url ...
# year
0 0 1 1 * pgslice add_partitions <table> --future 3 --url ...
Add a monitor to ensure partitions are being created.
SELECT 1 FROM
pg_catalog.pg_class c
INNER JOIN
pg_catalog.pg_namespace n ON n.oid = c.relnamespace
WHERE
c.relkind = 'r' AND
n.nspname = 'public' AND
c.relname = '<table>_' || to_char(NOW() + INTERVAL '3 days', 'YYYYMMDD')
-- for months, use to_char(NOW() + INTERVAL '3 months', 'YYYYMM')
-- for years, use to_char(NOW() + INTERVAL '3 years', 'YYYY')
Back up and drop older partitions each day, month, or year.
pg_dump -c -Fc -t <table>_202109 $PGSLICE_URL > <table>_202109.dump
psql -c "DROP TABLE <table>_202109" $PGSLICE_URL
If you use Amazon S3 for backups, s3cmd is a nice tool.
s3cmd put <table>_202109.dump s3:https://<s3-bucket>/<table>_202109.dump
Once a table is partitioned, make schema updates on the master table only (not partitions). This includes adding, removing, and modifying columns, as well as adding and removing indexes and foreign keys.
A few exceptions are:
- For Postgres 10, make index and foreign key updates on partitions only
- For Postgres < 10, make index and foreign key updates on the master table and all partitions
To undo prep (which will delete partitions), use:
pgslice unprep <table>
To undo swap, use:
pgslice unswap <table>
Set the tablespace when adding partitions
pgslice add_partitions <table> --tablespace fastspace
This set up allows you to read and write with the original table name with no knowledge it’s partitioned. However, there are a few things to be aware of.
When possible, queries should include the column you partition on to limit the number of partitions the database needs to check. For instance, if you partition on created_at
, try to include it in queries:
SELECT * FROM
visits
WHERE
user_id = 123 AND
-- for performance
created_at >= '2021-09-01' AND created_at < '2021-09-02'
For this to be effective, ensure constraint_exclusion
is set to partition
(default value) or on
.
SHOW constraint_exclusion;
Before Postgres 10, if you use INSERT
statements with a RETURNING
clause (as frameworks like Rails do), you’ll no longer receive the id of the newly inserted record(s) back. If you need this, you can either:
- Insert directly into the partition
- Get value before the insert with
SELECT nextval('sequence_name')
(for multiple rows, appendFROM generate_series(1, n)
)
For Postgres 10+, specify the primary key for partitioned models to ensure it’s returned.
class Visit < ApplicationRecord
self.primary_key = "id"
end
Before Postgres 10, preload the value.
class Visit < ApplicationRecord
before_create do
self.id ||= self.class.connection.select_all("SELECT nextval('#{self.class.sequence_name}')").first["nextval"]
end
end
Please submit a PR if additional configuration is needed.
You can also use pgslice to reduce the size of a table without partitioning by creating a new table, filling it with a subset of records, and swapping it in.
pgslice prep <table> --no-partition
pgslice fill <table> --where "id > 1000" # use any conditions
pgslice swap <table>
Triggers aren’t copied from the original table. You can set up triggers on the intermediate table if needed. Note that Postgres doesn’t support BEFORE / FOR EACH ROW
triggers on partitioned tables.
Postgres 10 introduces declarative partitioning. A major benefit is INSERT
statements with a RETURNING
clause work as expected. If you prefer to use trigger-based partitioning instead (not recommended), pass the --trigger-based
option to the prep
command.
Always make sure your connection is secure when connecting to a database over a network you don’t fully trust. Your best option is to connect over SSH or a VPN. Another option is to use sslmode=verify-full
. If you don’t do this, your database credentials can be compromised.
With Homebrew, you can use:
brew install ankane/brew/pgslice
Get the Docker image with:
docker pull ankane/pgslice
alias pgslice="docker run --rm -e PGSLICE_URL ankane/pgslice"
This will give you the pgslice
command.
If installation fails, your system may be missing Ruby or libpq.
On Mac, run:
brew install libpq
On Ubuntu, run:
sudo apt-get install ruby-dev libpq-dev build-essential
Run:
gem install pgslice
To use master, run:
gem install specific_install
gem specific_install https://github.com/ankane/pgslice.git
Also check out:
- Dexter - The automatic indexer for Postgres
- PgHero - A performance dashboard for Postgres
- pgsync - Sync Postgres data to your local machine
Everyone is encouraged to help improve this project. Here are a few ways you can help:
- Report bugs
- Fix bugs and submit pull requests
- Write, clarify, or fix documentation
- Suggest or add new features
To get started with development:
git clone https://github.com/ankane/pgslice.git
cd pgslice
bundle install
createdb pgslice_test
bundle exec rake test
To test against different versions of Postgres with Docker, use:
docker run -p=8000:5432 postgres:10
TZ=Etc/UTC PGSLICE_URL=postgres:https://postgres@localhost:8000/postgres bundle exec rake
On Mac, you must use Docker for Mac for the port mapping to localhost to work.