[meta] Backup/Restore improvements
We've been getting reports of backup/restore issues.
- A customer reporting that backing up can take >24hrs.
- A customer reporting that restore fails. This is due to a special RO dotfile in their environment that is automatically created in all directories due to a point-in-time backup/restore solution on the drive. The issue occurs when we try to copy out of the temporary directory during the restore process, these files get picked up by the copy.
- A customer reporting that restores can take a very long time, and chew up a 4 core box for hours. (Did not wait for it to complete)
- We've also gotten a bug report here with tar: https://gitlab.com/gitlab-org/gitlab-ce/issues/46891#note_95048966
Edited by Joshua Lambert