Unsuccessful restoration of a TASK from the backup

[AnSMru]

Actions before raising this issue

• I searched the existing issues and did not find anything similar.
• I read/searched the docs

Steps to Reproduce

After every unsuccessful restoration of a TASK from the backup, raw images are created in the docker cvat_server/~/data/data/number_job.
How can this be avoided? And how to ensure that the directory is deleted after a failed restoration to prevent disk space clutter?
I had corrupted image files (broken ones) that were previously backed up in version 1.92. Now, when attempting restoration in 2.31.1, multiple attempts were made, and each attempt created its own job directory, copied files from the archive, but ultimately resulted in an error.

Import Backup
Started by admin on Mar 24th 25, 15:02
rest_framework.exceptions.ValidationError: [ErrorDetail(string='Incorrect file mapping to manifest content', code='invalid')]

Expected Behavior

No response

Possible Solution

No response

Context

No response

Environment

[bsekachev]

Hello,

All data are expected to be deleted in case of unsuccessful restore, it is enforced by django signals mechanism. When corresponding Task/Data model removed from the database, server should remove corresponding data from the filesystem (data/data/<data_id>, data/tasks/<task_id).

May you provide a minimal backup file, reproducing the issue?

[AnSMru]

After exporting, there was a file like this in the archive. And accordingly, it cannot be read during import.