Status: Fixed
I found the error in the USN handling code.
The problem was that the filename was compared case insensitive, when doing USN file listing.
This caused all my tests to pass (as the folders have the correct case).
I managed to type it with the incorrect case in a test setup and found the error.
I will put up a new preview soon with the fix.
I would recommend that you use that version instead and delete the usn-policy setting from the backup.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@605 59da171f-624f-0410-aa54-27559c288bec
Status: Fixed
The implementation is a mix of the the three mentioned methods (issue #247).
There is a commandline option called --open-file-policy, which can be set to
"ignore", "snapshot" or "copy".
The "ignore" setting does the same as previous version, simply exclude the file from the
backup set.
If the setting is either "snapshot" or "copy", and the file it locked, it is opened in non-excusive read mode.
Once open, a in-memory signature file is generated. This is compared to the previous signature to detect changes,
just as the normal operation. If the file has not changed, the file is skipped as it would normally be.
If the file has changed and the policy is "snapshot", the file will be processed as normal,
but during the file read, a new signature is generated. This new signature is stored
in the signature archives. When writing the new signature, it is compared to the original signature.
If the file has changed during the backup operation, a warning is written to the output log.
If the policy was set to "copy", the modified file is copied to a temporary location,
while a new signature is generated. If the copy fails due to a file change, the file
is omitted from the backup set. If the copy succeeds, the file is known to not have changed
during the first access and the copy, which gives some assurance that the file is not
actively being written, at the expense of a copy operation.
The default setting is "snapshot", because this is likely to work in most scenarios,
and potentially produces a broken file rather than guarantee a missing file (by "ignore").
Note that when reading an open file, there is always a chance that the file is being written,
so you may end up with a broken file in the backup. This is also true for the VSS and LVM snapshots,
except for VSS-aware applications.
The "copy" method is meant to minimize this risk, but there is no guarantee that the file is in an
acceptable state, just because it has not been written for any period.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@588 59da171f-624f-0410-aa54-27559c288bec
Status: Fixed
USN is now activated by default on windows, but requires Administrative privileges.
If it fails to activate, it is silently ignored, and normal filesystem enumeration is used.
Set --usn-policy to:
"on" to get a warning if it fails,
"required" to abort the backup if it fails
"off" to disable USN
"auto" to silently ignore a fail
git-svn-id: https://duplicati.googlecode.com/svn/trunk@584 59da171f-624f-0410-aa54-27559c288bec
Added the --vss-exclude-writers option.
Update issue #260
Status: Fixed
I have now added a --vss-exclude-writers commandline option.
I had another look at the other supported options, and I consider this the most important.
I made the simplest possible version, so it only supports class guids, not component names
or writer instance ids.
If you need any of the other switches or the component names, please let me know.
I would also be interested in hearing that it actually works (I get no errors, but thats no proof).
You can specify multiple writers with a semicolon: --vss-exclude-writers="{d61d61c8-d73a-4eee-8cdd-f6f9786b7124};{0bada1de-01a9-4625-8278-69e735f39dd2}"
If you need the guids, you can get them with the following command in a command prompt:
vssadmin list writers
git-svn-id: https://duplicati.googlecode.com/svn/trunk@574 59da171f-624f-0410-aa54-27559c288bec
Fixed a protential race condition with warning messages and file counts (cosmetic only) when using async transfers.
Spelling correction, Seperator -> Separator.
Changed so a restore operation can also supply advanced options.
Changed the format of timestamps in filenames to follow the Duplicity convention, which removes the need for --time-separator and --use-short-filenames.
Moved the URL utility into Core, to avoid compilation errors with Monodevelop.
Reorganized the runner to better support the LiveControls on all operations rather than only for backup/restore.
Update issue #210
Status: Fixed
I have now changed the time format to follow the Duplicity naming convention.
This means that all filenames are now generated with the time set as "yyyyMMddTHHmmssZ".
This format is more portable across all platforms, and easily sortable/readable.
The drawback is that the Z indicates UTC time, so if one looks at the filenames,
the times are not local. The motivation for this is to avoid trouble
with daylight savings time offsets.
There is full backwards compatibility as long as backups have not used more than
one character for --time-separator (only possible via commandline).
This new format means that the options --time-separator and --use-short-filenames
are now deprecated and removed from the user interface. It is possible to
access them from the advanced options page, along with the --use-old-filenames option.
If anyone needs these options for some reason, let me know, otherwise they will be removed
in a later version.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@523 59da171f-624f-0410-aa54-27559c288bec
Added snapshot support using either VSS (win xp/vista/7) or LVM (linux).
Using the snapshot feature requires administrative/root privileges, and is toggled by the "snapshot-policy" option (by commandline or in advanced settings).
Also improved the validation of options to not show incorrect warnings.
This fixes issue #233.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@461 59da171f-624f-0410-aa54-27559c288bec
