The number of methods required by the ICompression interface is now reduce and all handling of path conversion and text encoding is now done in the CompressionWrapper class. This should make it easier to implement a new compression module, and reduce the number of errors that occur because of filename/encoding issues.
This commit also contains a fix for reporting a sane error message if the compression module is not found/loaded.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@1417 59da171f-624f-0410-aa54-27559c288bec
There is now an option called "vss-use-mapping" that activates DefineDosDevice and uses short DOS like paths.
I also find a little place that could be a performance problem, so maybe this fixes the issue even without the mapping.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@1413 59da171f-624f-0410-aa54-27559c288bec
This fixes issues #144, #577 and #320.
This update introduces a new options called "symlink-policy" that can be set to "store", "ignore" or "follow".
The default setting "store" means all symlinks are recorded as symlinks and recreated as symlinks on restore. This is chosen as default because it is the best replication of the actual disk contents.
Setting it to "ignore" simply means that no information is recorded about any symlinks.
And lastly, setting it to "follow" reverts to the previous default of treating symlinks as if they are normal files/folders.
Caveat: On Windows, you need Administrator privileges to create symlinks (when restoring), but not for reading them (on backup).
Since this change meant that the file attributes must be read, I also added support for the option "exclude-files-attributes", that can be used to ignore/exclude files with certain attributes, such as "hidden", "archive" or "encrypted".
Since I had to much about so much with getting the symlinks supported on Windows, I also added some extra logic to the Windows file handler, so it can now handle files and folders with up to 32k chars in the path. The logic is done so that the normal System.IO operations are used for paths with less than 260 characters (MAX_PATH), and on error or longer paths, the Win32 Unicode API is invoked. This should give maximal backwards compatibility and still allow backup and restore of files with long paths.
This commit also includes a BlockingQueue implementation that is supposed to make file/folder enumeration faster, but this has not been implemented because Duplicati depends on having a "whole world" view of the files and folders before running the backup (i.e. deeper changes required for this).
git-svn-id: https://duplicati.googlecode.com/svn/trunk@1381 59da171f-624f-0410-aa54-27559c288bec
Status: Fixed
The problem is that the the file-based backend returns "Directory not found" when asking for the not-yet-mounted folder.
Duplicati assumed that this error means that the folder will be created later, and thus returned an empty list,
as the folder would be empty when created. This means that the retry logic was not used.
Backing up to an empty folder will then result in a full backup being made.
But once the files started uploading, the folder is there and the folder is not created, but it is still a full backup.
This commit changes the logic a bit, so the list operation will not return until successful and a missing folder will be created during the list operation.
This ensures that the backup does not start before there has been a successful listing of the folder. If it fails, there is the normal retry delay, which should be enough to ensure that the folder is re-mounted. There may be warnings in the log, that the auto create failed. I have removed the logic that will auto create the folder during upload, as it should no longer be possible to get to the upload stage without verifying that the folder exists.
I have also changed the logic of the --disable-autocreate-folder option, so default value is "false" if it is a backup, and "true" for any other operation, so simply issuing a "list" or "cleanup" will not create folders (it did not do that before either).
git-svn-id: https://duplicati.googlecode.com/svn/trunk@1259 59da171f-624f-0410-aa54-27559c288bec
Status: Fixed
Duplicati now defaults to not upload a backup if nothing has changed.
This is indicated by a light-green check mark in the status window.
The option --upload-unchanged-backups can be set to revert to the previous way of working.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@958 59da171f-624f-0410-aa54-27559c288bec
Status: Fixed
There is now a delete transaction file being written, which should prevent cases where a large number of orphan files are detected.
If a large number of orphan files are detected, this should now be considered a real error.
I chose not to put in the cleanup button, because it would take too long for me to fit it in there, and I would rather focus on getting 2.0 ready, which will have a new UI anyway.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@957 59da171f-624f-0410-aa54-27559c288bec
Status: Fixed
Changed to the AWSSDK library, which fixes a few quirks, but basically seems to be the Affirma version slightly renamed.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@811 59da171f-624f-0410-aa54-27559c288bec
Status: Fixed
Thanks to Livio, this is now fixed and can be toggled with the --allow-sleep command.
It will be included in the next preview release.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@807 59da171f-624f-0410-aa54-27559c288bec
The code should now "gracefully" handle missing timestamps.
For files with no modification time, they are assigned 1. jan. 1970 00:00:00 (EPOCH).
This has the drawback that files are read fully because we cannot rely on the modification timestamp, so it may be a bit slower.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@802 59da171f-624f-0410-aa54-27559c288bec
Now also stores the filename+size of signature/content files to ensure that the parsed filenames match the expected ones.
The filesize and filenames are used to verify that the remote file is complete, without downloading it.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@716 59da171f-624f-0410-aa54-27559c288bec
This commit includes an option to create an unencrypted verification file.
There is now a check to see if all volumes mentioned in the manifest exists,
before performing backup/restore operations.
This check is also performed on cleanup.
There is now a check that prevents auto-cleanup from removing more than
2 orphan files.
If a delete-older-than or delete-all-but-n-full is interrupted, the leftover
files will not be auto-deleted but must be deleted with the cleanup command,
which allows the user to see what files are delete.
Update issue #360
Status: Fixed
There is now a complete verification chain, where the manifest hashes are verified backwards.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@711 59da171f-624f-0410-aa54-27559c288bec
Status: Fixed
Now allows 1% (max 1 hour) slack in the time difference.
Can be disabled with the --disable-time-tolerance flag.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@595 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
Status: Fixed
Added a find-last-version command to the commandline version,
it takes the same input as a restore operation, so
use --file-to-restore to supply a list of files to search for.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@559 59da171f-624f-0410-aa54-27559c288bec
Previously some backends autocreated a missing folder, and some threw an exception.
The new standard is that all backends throw the FolderMissingException on all access (except delete).
This means that the Test() and List() methods are now the same for all stock backends.
Third-party applications that use the backends should implement custom logic to auto-create the folders,
when the FolderMissingException is encountered.
Duplicati auto-creates missing folders, but this can be toggled through the new flag --disable-autocreate-folders.
In case the auto-create fails to detect a missing folder, there is now a "create-folder" action in the commandline version.
Update issue #102
Status: Fixed
The IBackend_v2 interface has the CreateFolder and Test methods.
The CreateFolder is exposed to the commandline interface with the action create-folder.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@558 59da171f-624f-0410-aa54-27559c288bec
Added name support in GUI and commandline version.
When restoring from commandline, subfolders will now be named like the source folders, if only one source folder is given.
When restoring from GUI, the names will be suggested like in the commandline.
It is no longer required that the unused folders exists when restoring a multi-source backup from the GUI.
Added size tooltips to the restore GUI window with the list of backup chains.
Now correctly deletes old signature cache files, even if they use the old names.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@542 59da171f-624f-0410-aa54-27559c288bec
Update issue #235
I have now added a simple version of the collection-status command,
which shows Type, Time, Volume count and Size for each backup set,
with the incrementals listed under the full's they belong to.
git-svn-id: https://duplicati.googlecode.com/svn/trunk@540 59da171f-624f-0410-aa54-27559c288bec
Status: Fixed
A restore now defaults to background and some filenames are also shown while restoring.
The "direct restore" option cannot run in the background as it requires the schedule to be stored in the database.
This can be fixed after version 2.0
git-svn-id: https://duplicati.googlecode.com/svn/trunk@525 59da171f-624f-0410-aa54-27559c288bec
