~germar/duplicity/par2

  • Committer: Kenneth Loafman
  • Date: 2013-01-08 01:28:17 UTC
  • mfrom: (902.1.9 vol-corruption)
  • Revision ID: kenneth@loafman.com-20130108012817-h0qolw69tfkncby4
* Merged in lp:~mterry/duplicity/static-corruption
  - This branch fixes three possible ways a backup could get data-corrupted.
    Inspired by bug 1091269.
      A) If resuming after a volume that ended in a one-block file, we would
         skip the first block of the next file.
      B) If resuming after a volume that ended in a multi-block file, we would
         skip the first block of the next file.
      C) If resuming after a volume that spanned a multi-block file, we would
         skip some data inside the file.
  - A and B are because when finding the right place in the source files to
    restart the backup, the iteration loop didn't handle None block numbers
    very well (which are used to indicate the end of a file).
  - C is what bug 1091269 talks about. This was because data block sizes would
    get smaller as the difftar file got closer and closer to the volsize.
    Standard block sizes were 64 * 1024.  But say we were close to the end of
    the difftar... When resuming, duplicity doesn't know the custom block sizes
    used by the previous run, so it uses standard block sizes. And it doesn't
    always match up, as you can imagine. So we would leave chunks of data out
    of the backed up file.
  - Tests added for these cases.
  - This branch is called 'static-corruption' because all these issues occur
    even when the source data doesn't change. I still think there are some
    corruption issues when a file changes in between duplicity runs. I haven't
    started looking into that yet, but that's next on my list.
  - C only happened without encryption (because the gpg writer function already
    happened to force a constant data block size). A and B happened with or
    without encryption.
Filename Latest Rev Last Changed Committer Comment Size
..
bin 800.1.24 12 years ago Michael Terry Make tarball layout match bzr layout much more clo Diff
dist 1 21 years ago bescoto Initial checkin Diff
duplicity 1 21 years ago bescoto Initial checkin Diff
po 355 15 years ago loafman patch #6700: Make duplicity translatable https://s Diff
testing 1 21 years ago bescoto Initial checkin Diff
.bzrignore 804 12 years ago Kenneth Loafman Remove random_seed from VCS, adjust .bzrignore. 71 bytes Diff Download File
CHANGELOG 903 11 years ago Kenneth Loafman * Merged in lp:~mterry/duplicity/static-corruption 71.6 KB Diff Download File
Changelog.GNU 903 11 years ago Kenneth Loafman * Merged in lp:~mterry/duplicity/static-corruption 162 KB Diff Download File
COPYING 476 15 years ago loafman After email voting among known duplicity contribut 17.5 KB Diff Download File
README 875.1.2 11 years ago ede some clarifications in README 2.6 KB Diff Download File
README-LOG 813 12 years ago Kenneth Loafman Some doc changes including new requirements. 1.1 KB Diff Download File
README-REPO 813 12 years ago Kenneth Loafman Some doc changes including new requirements. 809 bytes Diff Download File
tarfile-CHANGES 777.1.8 13 years ago Michael Terry use python2.7's tarfile instead of whichever versi 102 bytes Diff Download File
tarfile-LICENSE 783 12 years ago Kenneth Loafman Merged in lp:~mterry/duplicity/tarfile 2.7 KB Diff Download File