Version 3.0.5 of xfsdump has been released.
A gzipped tar archive of the source code is available here:
The source code can be accessed via git using this URL:
Below is a summary (from the doc/CHANGES file) of changes
since release 3.0.4:
xfsdump-3.0.5 (30 March 2011)
- Release tags will now be digitally signed
- Quota files will now be dumped, regardless of the maxsize
- The new "-s sessid" flag allows inventory sessions to be
pruned by their session ID.
- Fixed a bug in handling long dump filenames, and dropped
(undocumented) support for encoding certain parameters
within the dump filename
- NODECHK is now off by default, meaning xfsrestore will now
support 16 times more directory entries (4 billion)
- nrh_t is now 64 bits wide, allowing xfsrestore to support
dumps with up to 4 billion directory entries
- nix_t is no longer useful, and has been eliminated
- Memory use in xfsrestore is better managed now. Segments
of nodes are now power-of-2 sized, and allocated nodes are
no longer needlessly zeroed and linked into the free list.
- Pathname resolution in xfsrestore has a number of
- Better checking has been implemented for compatibility
when resuming a cumulative restore
- Build system output has been cleaned up considerably
- Dead "namreg" code has been removed
- Build dependencies are now determined automatically
- Other miscellaneous build system improvements, as well bug
fixes thanks to Mike Frysinger and Arkadiusz Miśkiewicz.