PGRestore from S3

A small utility to restore a database from a dump-file (gzipped or not) in S3. Companion to PGDump to S3.

Requirements

Set these env-vars for pg_rearoew (see PG ENV-vars):

• PGHOST
• PGPORT (unless 5432)
• PGUSER
• PGPASSWORD (and if need be, other PG env vars according to your needs)

and these for AWS CLI (or less if using EC2-roles) (see AWS CLI ENV-vars):

• AWS_ACCESS_KEY_ID
• AWS_SECRET_ACCESS_KEY
• AWS_DEFAULT_REGION

Temp-volume

Be sure to bind a temporary volume to /restoretemp (or somewhere else, but set env TEMP_RESTORE_DIR to that).

Command

Entrypoint is the script restore_from_s3.

The script takes these parameters: s3://mybucket/myfolder/mydump target_database

If dump has .gz suffix, it will ungzip-it.

pg_restore parameters

Utility will use these parameters for restore: pg_restore --clean --no-acl --no-owner -d target_database mydump.

Example

docker run --rm -it -e PGHOST=xx -e PGUSER=xx -e PGPASSWORD=xx -e AWS_DEFAULT_REGION=eu-west-1 -e AWS_ACCESS_KEY_ID=xx -e AWS_SECRET_ACCESS_KEY=xx -v /restoretemp upptec/pgrestore_from_s3 s3://my_bucket/my_folder/mydump.gz mydb

Howto dump

See PGDump to S3.