Use bulk insertion to Postgres at startup to reduce time taken for initial snapshot restore for postgres plugin. Avoid duplicate writes of accounts at startup. Doing account plugin notification and indexing in parallel.
Improved error handling for postgres plugin to show the real db issues for debug purpose
Added more metrics for postgres plugin.
Refactored plugin centric code out to a sub module from accounts_db and added unit tests
* add filler accounts to bloat validator and predict failure
* assert no accounts match filler
* cleanup magic numbers
* panic if can't load from snapshot with filler accounts specified
* some renames
* renames
* into_par_iter
* clean filler accts, too
Summary of Changes
Create a plugin mechanism in the accounts update path so that accounts data can be streamed out to external data stores (be it Kafka or Postgres). The plugin mechanism allows
Data stores of connection strings/credentials to be configured,
Accounts with patterns to be streamed
PostgreSQL implementation of the streaming for different destination stores to be plugged in.
The code comprises 4 major parts:
accountsdb-plugin-intf: defines the plugin interface which concrete plugin should implement.
accountsdb-plugin-manager: manages the load/unload of plugins and provide interfaces which the validator can notify of accounts update to plugins.
accountsdb-plugin-postgres: the concrete plugin implementation for PostgreSQL
The validator integrations: updated streamed right after snapshot restore and after account update from transaction processing or other real updates.
The plugin is optionally loaded on demand by new validator CLI argument -- there is no impact if the plugin is not loaded.
When reconstructing the AccountsDb, if the storages came from full and
incremental snapshots generated on different nodes, it's possible that
the AppendVec IDs could overlap/have duplicates, which would cause the
reconstruction to fail.
This commit handles this issue by unconditionally remapping the
AppendVec ID for every AppendVec.
Fixes#17088
* Return error if Transaction locks an executable as writable
* Return error if a ProgramData account is writable but the upgradable loader isn't present
* Remove unreachable clause
* Fixup bpf tests
* Review comments
* Add new TransactionError
* Disallow writes to any upgradeable-loader account when loader not present; remove is_upgradeable_loader_present exception for all other executables
Jeff Washington (jwash)
Lijun Wang
Michael Vines
Brooks Prumo
Tyera Eulberg