v0.1.7
This release fixes issue #63. This issue led to the possibility of an incorrect point of restoration when the restoration was done without "standby_mode = on". The issue was reproducible only on Amazon EBS drives and frequency of reproduction varied, thanks @bartdag for extensive help with researching this issue.
Also, this release brings features shown necessary by QA cycle on the heavily loaded clusters:
- Support of parallel
wal-push
with reading from archive_status directory (relates to #58 ) - WAL-G was failing during
backup-push
if files were disappearing during tar creation. This was fixed (relates to #65 ) - Concurrency of
wal-prefetch
was improved
Some edge-case error handling and code quality work was done by @ferhatelmas. Thanks to @benubois and @lesovsky for reporting problems and sharing their experience.