FreeBSD fixes multiple issues in hastd replication service
Key Takeaway
Multiple vulnerabilities and bugs in the hastd service can lead to data leakage in metadata or prevent storage replication.
Summary
The FreeBSD Project has released security advisory FreeBSD-EN-26 addressing three distinct issues in the hastd(8) service. These issues affect all supported versions of FreeBSD and can result in uninitialized memory being written to storage metadata, command failures in hastctl, or nodes becoming stuck in a degraded state during replication.
Technical breakdown of hastd issues
- Information Leakage: The nvpair implementation fails to initialize padding, allowing small amounts of data from previous heap allocations (such as configuration file text or previous resource data) to be embedded in HAST resource metadata or network messages.
- Control Failures: An assertion failure occurs when hastctl sends empty control messages to hastd, causing commands to abort.
- Replication Hangs: A bug in how worker processes read socket types via local domain sockets can cause secondary nodes to hang indefinitely during initialization, preventing the system from transitioning out of degraded mode.
Why it matters
Users relying on HAST for transparent block storage replication should update immediately to prevent potential data exposure in metadata or loss of redundancy.
Original reference: www.freebsd.org →