- Instead of very small (5 MB-ish) segment files, use larger ones
- Request asynchronous write-out or write-through (TODO) where it is supported,
to achieve a continuously high throughput for writes
- Instead of depending on ordered writes (write data, commit tag, sync)
for consistency, do a double-sync commit as more serious RDBMS also do
i.e. write data, sync, write commit tag, sync
Since commits are very expensive in Borg at the moment this makes no
difference performance-wise.
New platform APIs: SyncFile, sync_dir
[x] Naive implementation (equivalent to what Borg did before)
[x] Linux implementation
[ ] Windows implementation
[-] OSX implementation (F_FULLSYNC)
Update installation instructions so that the following error won't occur:
gcc: error: /usr/lib/rpm/redhat/redhat-hardened-cc1: No such file or directory
error: command 'gcc' failed with exit status 1
/mnt/backup was confusing as people like to mount their backup disk on /mnt/backup,
but borg init /mnt/backup does not work if that directory already exists because it is
the mountpoint. it would work, if /mnt was the mountpoint, but that is not obvious
and also unusual.
Use with caution: permanent data loss by specifying incorrect patterns
is easily possible. Make a dry run to make sure you got everything right.
borg recreate has many uses:
- Can selectively remove files/dirs from old archives, e.g. to free
space or purging picturarum biggus dickus from history
- Recompress data
- Rechunkify data, to have upgraded Attic / Borg 0.xx archives deduplicate
with Borg 1.x archives. (Or to experiment with chunker-params for
specific use cases
It is interrupt- and resumable.
Chunks are not freed on-the-fly.
Rationale:
Makes only sense when rechunkifying, but logic on which new chunks to
free what input chunks is complicated and *very* delicate.
Future TODOs:
- Refactor tests using py.test fixtures
-- would require porting ArchiverTestCase to py.test: many changes,
this changeset is already borderline too large.
- Possibly add a --target option to not replace the source archive
-- with the target possibly in another Repo
(better than "cp" due to full integrity checking, and deduplication
at the target)
- Detect and skip (unless --always-recompress) already recompressed chunks
Fixes#787#686#630#70 (and probably some I overlooked)
Also see #757 and #770
- Group options
- Nicer list of options in Sphinx
- Deduplicate 'Common options'
(including --help)
The latter is done by explicitly declaring --help in the common_parser,
which is then inherited by the sub-parsers; no change in observable
behaviour.
ubuntu was showing up twice in the list of supported OSes... it seems it was because the line was getting too long, so I removed the "names" and kept only the numbers to keep the line short.
