Releases: gboudreau/Greyhole
Releases · gboudreau/Greyhole
0.9.14
- New feature: --fsck --checksums
- New feature: --fsck --delete-orphaned-metadata
- Bugfix: prevent Greyhole from writing metadata files in directories where a mount disappeared after the current --fsck started
- Bugfix: prevent an infinite loop that could occur when creating new directories
- Bugfix: some fsck details were missing from emailed report
- Bugfix: better balancing during a drive replacement fsck
- Bugfix: during fsck, don't treat temporary files as real orphaned files, just delete them
- Bugfix: quiet greyhole-dfree warnings when some storage pool directories became unavailable
- Bugfix: USAGE file asked user to make a copy of greyhole.example.conf; this is already taken care of in INSTALL and in RPM/DEB packages
- Improved logging when creating new file copies (during fsck and normally)
- Improvement: Now using cp instead of rsync by default for file copies
- Improvement: settings are now backed up to the data drives, and automatically restore when needed
- Improvement: new recommendations on dir_selection_algorithm
- Improvement: you can now use the keyword 'max' in greyhole.conf num_copies settings
- Improvement: Outputting more details on RPM installs, to try to identify the problem with the missing symlinks on Amahi's hda-greyhole updates
- Improvement: make sure all fsck_file tasks have been executed before running another fsck
- Improvement: will read more spooled tasks at once, instead of wasting time listing again and again the spool directory
- Improvement: Better --status report
- Improvement: increased max queued tasks default for MySQL from 100K to 10M
- Improvement: fixed double-/ typos in log
- Improvement: various code refactoring, quiet warnings, etc.
- Improvement: more verbose PHP warnings in log
- Improvement: faster MySQL query to find the next tasks (added correct index)
- Improvement: better looking --view-queue
- Improvement: --status now detects when the daemon is stopped
- Improvement: better looking logs when multiple threads are doing logging simultaneously
- Improvement: simplified code that checks & mails fsck logs
- New build scripts; cleaner and more automated than the previous uncommitted scripts I used to create new builds
- Committing logo PSD
- Clean up: remove --postrotate and --prerotate commands
- Minor cosmetic changes in the code
0.9.13
0.9.12
0.9.11
0.9.10
- New options to prioritize specific drives: see https://github.com/gboudreau/Greyhole/blob/master/greyhole.example.conf, Storage Selection Algorithm section
- Issue #20 - Bugfix: renaming then deleting a folder full of files on a big share, GH would take forever to process the rename operation!
- Issue #6 - Cent-OS (5.6) compatibility, and documentation clarifications
- Issue #24 - Change greyhole terminology
- Issue #28 - 32bit int truncation for large drives
- Bugfix: Quiet warning when no dir_selection_groups are defined in greyhole.conf
- Fixing --balance; would skip too many shares when sticky files are used. Also display more information in the log when a file isn't moved, to help debug why that is.
- More --balance fixes, this time to support both drives that need to be filled, and those that need to be freed.
- Small changes to USAGE and UPGRADING instructions
- Quiet annoying "Optimizing MySQL tables..." log when using greyhole from the command line
- Quiet warning in greyhole.log
- Improvement: allow a trailing / in --going, --gone, --wait-for and --thaw commands.
- Removed stray var_dump!
- Improvement: USAGE: "follow symlinks = yes" is not needed in smb.conf