Evidence collection
Release notes
HAF Release Notes
-
#210: Removed
row_id
indexes for irreversible data contexts. -
Tables and views owned by application contexts are now placed in separate schemas, allowing the same application to run simultaneously in the same HAF database.
-
Encoded operation type and block number in the operation ID value to reduce disk space usage in the
operation
table. -
Introduced a new, much simpler method for writing application loops that encapsulates context attachment, transaction committing, and explicit application state definitions.
-
Moved non-updatable elements of HAF to the
hafd
schema, while updatable elements remain inhive
. -
#244: Automatically log application state changes.
-
#249: Support for applications to create indexes.
-
Possible to update state providers.
-
Use
pg_cron
to execute context detachment for dead applications. -
Added
keyauth
state provider.
Several bug fixes and Infrastructure updates, including dockerization, enhanced test coverage, CI/CD pipeline improvements
Evidence collection
Release notes
Warning: this update (due to underlying Hived modifications) changes hived/HAF state format and requires HAF node replay, if you are using version 1.27.4.0 or earlier.
New features, functional improvements, and optimizations
- Support for the new Query Supervisor (limits database resources available to database roles during a query). See documentation located here: README.md
- Dockerized environment supports --load-snapshot/--dump-snapshot options backup/restore a HAF server (including hived): !282
- Optimization of
hive.operation
type conversion into textual format (improves backup/restore times): !292 . Right now by defaulthive.operation
is converted intobytea
literal (representing operation binary format common to blockchain storage). To use the legacy JSON-textual form, use thebody
column provided byhive.operations_view
(also in app-specific operations view) or change your app to use the newbody_binary
column (this will improve performance).
- New function available for calculating HivePower value: !287
- Complex HAF apps can now be built from multiple smaller apps: !277
Bugfixes:
- Fixed a constraint violation while processing data in Account Metadata State Provider (common Metadata Collector provided by HAF, next to be instantiated at HAF-app level): !274
- Eliminated hang of hived node while closing after earlier caught PostgreSQL related error: !278
- Community reported problem related to processing forks by the HAF server: !285
- Multiple app roles are allowed to create app-specific tables like also perform synchronization (simplifies and makes deployments much more secure):
- HAF functions will explicitly fail when app-specific table identifiers are too long: !295
Evidence collection
Release notes
Warning: this update changes hived/HAF state format and requires HAF node replay, if you are using version 1.27.3 or earlier.
- . New features:
- Optimized HAF operation storage to reduce its size (Hive operations are now serialized to binary form in same way as blockchain does): !201
- Optimized Hive operation to JSONB type conversion: !228
- HAF knows hardfork data and stores this info in the dedicated table: !172
- HAF provides subset of basic DGPO data specific to given block: !213
- HAF is able to store and verify a database schema to match running HAF node to existing database: !229
- Support for dump and load HAF instances (specific to backup administration purposes): !194
- Prepared Account Metadata Consensus Data State provider to be used at implementation of account metadata HAF application
- Preliminary boilerplate for building and deploying HAF applications using docker compose: !217
- Upgrade to Ubuntu 22.04 as a base supported OS
- Devops/automations:
- Bugfixes:
- fixed deadlock during joining massive operation processor
Evidence collection
Release notes
First release of Hive Application Framework, based on Hived version: v1.27.3