v2.1.1

This is a release of the v2 schema with the addition of the new eth_meta schema and eth_meta.watched_addresses table.

Changelog: release-v2.0.0...release-v2.1.1

v3.1.0

v3 Database Schema

The primary objective of this upgrade from v2 is to move from a serial int based primary/foreign key scheme to a deterministic natural key scheme, this helps facilitate horizontal scaling by avoiding serial key conflicts when merging separate databases.

We also introduce sets of migrations split into pre- and post- historical batch processing units. The pre- set forgoes all FK and unique constraints, indexes, and sets the tables as UNLOGGED. The post- set adds these constraints and indexes and sets the tables as LOGGED. This is to improve performance of historical batch processing.

Additionally,

• New watched addresses table
• Remove unique constraint on the receipt_cids.leaf_mh_key column, since we can (in theory) have duplicate receipts
• Adding some additional rows that will be useful for searches (e.g. header_cids.coinbase).
• Removing some that are not particularly useful for searches (e.g. header_cids.base_fee).
• Switching some BIGINT types to NUMERIC to avoid overflow.
• Typo fixes in file names.
• Split PK application to public.blocks into its own migration in the migrate_post_batch_set of migrations as we need to apply only that constraint prior to the v2 => v3 transformation.
• Split application of log_cids related indexes into separate migration for logTrie fix processing; disable afterwards for v2 => v3 transformation.
• Adjust migrations to (temporarily) apply log_cids.leaf_mh_key FK to ensure the logTrie fix was successful.
• New Makefile targets for running up migrations one at a time: migrate_up_by_one and migrate_post_batch_set_up_by_one.
• Fix issue where Makefile was failing to find goose if GOPATH was unset (#38).
• New table to track the current version of the database schema, the version corresponds with the semantic versioning of this release.

Full v2 => v3 changelog: release-v2.0.0...release-v3.1.0

v3.0.7

v3 Database Schema

The primary objective of this upgrade from v2 is to move from a serial int based primary/foreign key scheme to a deterministic natural key scheme, this helps facilitate horizontal scaling by avoiding serial key conflicts when merging separate databases.

We also introduce sets of migrations split into pre- and post- historical batch processing units. The pre- set forgoes all FK and unique constraints, indexes, and sets the tables as UNLOGGED. The post- set adds these constraints and indexes and sets the tables as LOGGED. This is to improve performance of historical batch processing.

Additionally,

• Remove unique constraint on the receipt_cids.leaf_mh_key column, since we can (in theory) have duplicate receipts
• Adding some additional rows that will be useful for searches (e.g. header_cids.coinbase).
• Removing some that are not particularly useful for searches (e.g. header_cids.base_fee).
• Switching some BIGINT types to NUMERIC to avoid overflow.
• Typo fixes in file names.
• Split PK application to public.blocks into its own migration in the migrate_post_batch_set of migrations as we need to apply only that constraint prior to the v2 => v3 transformation.
• Split application of log_cids related indexes into separate migration for logTrie fix processing; disable afterwards for v2 => v3 transformation.
• Adjust migrations to (temporarily) apply log_cids.leaf_mh_key FK to ensure the logTrie fix was successful.
• New Makefile targets for running up migrations one at a time: migrate_up_by_one and migrate_post_batch_set_up_by_one.
• Fix issue where Makefile was failing to find goose if GOPATH was unset (#38).
• New table to track the current version of the database schema, the version corresponds with the semantic versioning of this release.

Full v2 => v3 changelog: release-v2.0.0...release-v3.0.7

v3.0.6

v3 Database Schema

The primary objective of this upgrade from v2 is to move from a serial int based primary/foreign key scheme to a deterministic natural key scheme, this helps facilitate horizontal scaling by avoiding serial key conflicts when merging separate databases.

We also introduce sets of migrations split into pre- and post- historical batch processing units. The pre- set forgoes all FK and unique constraints, indexes, and sets the tables as UNLOGGED. The post- set adds these constraints and indexes and sets the tables as LOGGED. This is to improve performance of historical batch processing.

Additionally,

• Adding some additional rows that will be useful for searches (e.g. header_cids.coinbase).
• Removing some that are not particularly useful for searches (e.g. header_cids.base_fee).
• Switching some BIGINT types to NUMERIC to avoid overflow.
• Typo fixes in file names.
• Split PK application to public.blocks into its own migration in the migrate_post_batch_set of migrations as we need to apply only that constraint prior to the v2 => v3 transformation.
• Split application of log_cids related indexes into separate migration for logTrie fix processing; disable afterwards for v2 => v3 transformation.
• Adjust migrations to (temporarily) apply log_cids.leaf_mh_key FK to ensure the logTrie fix was successful.
• New Makefile targets for running up migrations one at a time: migrate_up_by_one and migrate_post_batch_set_up_by_one.
• Fix issue where Makefile was failing to find goose if GOPATH was unset (#38).
• New table to track the current version of the database schema, the version corresponds with the semantic versioning of this release.

Full v2 => v3 changelog: release-v2.0.0...release-v3.0.6

v3.0.0

v3 Database Schema

The primary objective of this upgrade from v2 is to move from a serial int based primary/foreign key scheme to a deterministic natural key scheme, this helps facilitate horizontal scaling by avoiding serial key conflicts when merging separate databases.

We also introduce sets of migrations split into pre- and post- historical batch processing units. The pre- set forgoes all FK and unique constraints, indexes, and sets the tables as UNLOGGED. The post- set adds these constraints and indexes and sets the tables as LOGGED. This is to improve performance of historical batch processing.

Additionally,

• Adding some additional rows that will be useful for searches (e.g. header_cids.coinbase)
• Removing some that are not particularly useful for searches (e.g. header_cids.base_fee)
• Switching some BIGINT types to NUMERIC to avoid overflow
• Typo fixes in file names.
• Split PK application to public.blocks into its own migration in the migrate_post_batch_set of migrations as we need to apply only that constraint prior to the v2 => v3 transformation.
• New Makefile targets for running up migrations one at a time: migrate_up_by_one and migrate_post_batch_set_up_by_one.
• Fix issue where Makefile was failing to find goose if GOPATH was unset (#38).
• New table to track the current version of the database schema, the version corresponds with the semantic versioning of this release.

Full v2 => v3 changelog: release-v2.0.0...release-v3.0.0

v0.3.3

This release contains some minor adjustments

• Typo fixes in file names.
• Split PK application to public.blocks into its own migration in the migrate_post_batch_set of migrations as we need to apply only that constraint prior to the v2 => v3 transformation.
• New Makefile targets for running up migrations one at a time: migrate_up_by_one and migrate_post_batch_set_up_by_one.
• Fix issue where Makefile was failing to find goose if GOPATH was unset (#38).

This is a patch version bump as there are no changes to the schema state and backwards compatibility with previous v0.3.x schemas is not broken.

v0.3.2

In this release a new table is added to track the current version of the database schema, the version corresponds with the semantic versioning of this release. This is a patch version bump as there are no breaking changes to the existing tables and backwards compatibility with previous v0.3.x schemas is not broken.

v0.3.1

New database schema, this is breaking upgrade so we are moving to a new minor version.

The primary objective of this upgrade is to move from a serial int based primary/foreign key scheme to a deterministic natural key scheme, this helps facilitate horizontal scaling by avoiding serial key conflicts when merging separate databases.

We also introduce sets of migrations split into pre- and post- historical batch processing units. The pre- set forgoes all FK and unique constraints, indexes, and sets the tables as UNLOGGED. The post- set adds these constraints and indexes and sets the tables as LOGGED. This is to improve performance of historical batch processing.

There are also some other smalls adjustments to the schema: switching some BIGINT types to NUMERIC to avoid overflow, adding some additional rows that will be useful for searches (e.g. header_cids.coinbase) and removing some that are not particularly useful (e.g. header_cids.base_fee).

Changelog: release-v0.2.1...release-v0.3.1

v0.3.0

New database schema, this is breaking upgrade so we are moving to a new minor version.

The primary objective of this upgrade is to move from a serial int based primary/foreign key scheme to a deterministic natural key scheme, this helps facilitate horizontal scaling by avoiding serial key conflicts when merging separate databases.

We also introduce sets of migrations split into pre- and post- historical batch processing units. The pre- set forgoes all FK and unique constraints, indexes, and sets the tables as UNLOGGED. The post- set adds these constraints and indexes and sets the tables as LOGGED. This is to improve performance of historical batch processing.

There are also some other smalls adjustments to the schema: switching some BIGINT types to NUMERIC to avoid overflow, adding some additional rows that will be useful for searches (e.g. header_cids.coinbase) and removing some that are not particularly useful (e.g. header_cids.base_fee).

Changelog: release-v0.2.1...release-v0.3.0

v2.0.0

This is a re-release of the v0.2.1 release, under the new major versioning scheme.

Changelog: release-v0.2.0...release-v2.0.0