The option is turned into an enum with values True, False, ListIndexes
True and False are for compatibility with the existing backups.
ListIndexes is equivalent to True but don't handle block volumes.
This is to make more workable a strategy of full remote testing
after each backup.
Before the change, it could happen that when the postbackup test
was finding problems in the backend, the job would be reported
as having completed successfully, the found problems being displayed
only in the complete log. After the change, an error is logged
that cause the job to be failed.
To repro, the following procedure can be used in a test backup:
change a hash for a block in the block table database and run a backup having set
--full-remote-verification and --full-block-verification to true and
--backup-test-percentage to 100.
* experimental faster db rebuild (#4955)
* try run update db query for every block file only when needed
add repair-force-block-use option
more appropriate naming for some routines
more tracing
more comments
Before the dindex files are deleted on the remote, update the state in the database to Deleting. Previously, only dblock files were handled this way.
Patch by warwickmm
Closes#4129
Direct import of PR#4978
Update web UI for new result reports.
For operations with fatal errors, write logs to same operation ID.
Test that Interrupted flag is correct in RunScriptTests.
Update backup log display for new result reporting.
Hide file statistics for fatal errors and change fatal icon.
This avoids creating directory %LocalAppData%\Duplicati when it would just be left
empty, and it makes running portable mode (e.g. debug builds) stealthier.
Add Storj DCS backend and deprecated Tardigrade DCS backend.
Storj renamed their "Tardigrade Decentralized Cloud Storage" product to "Storj DCS". This adds a new Storj DCS backend and deprecates the old Tardigrade one. A warning is issued to inform Tardigrade users to modify their configurations to use the new backend.
This also adds a passphrase verification field to ensure that the user has provided the correct encryption passphrase.
Previously, we would catch any Exception thrown by VerifyRemoteList and
perform an auto-cleanup (if specified in the options). However, we
should only perform the auto-cleanup if VerifyRemoteList detected an
issue with the uploaded files. Otherwise, an unrelated exception can
cause the database repair to be performed unnecessarily. In the case of
a connection issue, this can leave the database in a corrupted state.
This fixes#4516.
If we do not use an AS clause, the name of the column in the result set
is unspecified and may change depending on the version of SQLite.
https://www.sqlite.org/c3ref/column_name.html
This fixes#4403.
If there is more than one missing blocklist hash, we must reset the
offset index before copying hashes from the next blocklist into the
buffer.
This fixes#4397.
Since the GetSymlinkTarget method can throw an exception, we should log
the failure case and avoid further processing of the symlink target.
This concerns #2171, which describes issues with symlinks created using
Windows Subsystem for Linux.
