IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
Implementation overview:
- Incremental points are saved as dedicated rbd snapshots under the
"one_backup_<increment_id>" namespace. This snapshots are used to
generate delta files in rbdiff format.
- The rbdiff formats are stored in the backup server to restore the rbd volumes.
- The restore process is performed directly on the Ceph cluster
importing the base image (first full backup in the chain, rbd import)
and then applying the increments (rbd import-diff) up to the target
increment.
- Two new pseudo-protocols has been implemented to adopt the restore
pattern above (restic+rbd, rsync+rbd). This protocols bundle of the
rbdiff files in a tarball for transfer from the backup server. Note:
reconstruct process uses the Ceph BRIDGE_LIST and not the backup
server (as opposed to qcow2 backups)
Other bug fixes
- This commit also fixes#6741, resetting the backup chain after a
restore
- The original ceph drivers do not receive the full action information,
this now has been fixed by including VM information in the STDIN string sent
to the driver.
Compatibility note.
- backup actions should return now the backup format used raw, rbd, ...
If not provided oned (6.10.x) will use raw as a default to accommodate any
third party driver implementation. It is recommended to include this
third argument.
Signed-off-by: Guillermo Ramos <gramos@opennebula.io>
Co-authored-by: Guillermo Ramos <gramos@opennebula.io>
5f7b370c52