Restores and attachments
Restoring backups, and attaching them read-only.
attachment create
Attach one disk of a backup read-only to a running instance.
The point is to get at a few files without rehydrating the whole image. The attachment detaches and cleans itself up when its --ttl runs out.
openstack backup attachment create --point <point> --instance <instance> \
[--disk <resource-id>] [--device <device>] [--ttl <duration>] [--wait]
| Argument | Description |
|---|---|
--point <point> | Point to attach. Required. |
--instance <instance> | Running instance to attach it to. Required. |
--disk <resource-id> | Which disk of a multi-disk point to attach, by resource id; default the root disk. |
--device <device> | Guest device name, e.g. /dev/vdb. |
--ttl <duration> | Detach and clean up after this long, e.g. 30m or 24h (default: 4h) |
--wait | Wait for the attachment to be ready. |
attachment delete
Detach one or more attached backups.
openstack backup attachment delete <attachment> [<attachment> ...]
| Argument | Description |
|---|---|
<attachment> | Takes one or more. |
attachment list
List attached backups.
openstack backup attachment list [--point <point>] [--limit <limit>] \
[--project <project>]
| Argument | Description |
|---|---|
--point <point> | |
--limit <limit> | |
--project <project> | Another project's attachments (operator; needs a system-scoped token) |
restore create
Start a restore of a point.
volume and instance create a new resource, and only they can give
it a different shape: --disks for either, --flavor,
--networks, --key-name and --security-groups for instance
alone. bucket restores an object namespace into a new bucket.
in_place overwrites an existing resource, which destroys whatever is
currently there — it is the one mode that can lose data if aimed at the
wrong point, and is never a default.
openstack backup restore create --point <point> [--mode <mode>] \
[--name <name>] [--volume-type <type>] [--availability-zone <az>] \
[--tier <tier>] [--flavor <flavor>] [--networks <network,...>] \
[--key-name <key-pair>] [--security-groups <group,...>] \
[--disks <resource-id,...>] [--target-project <project>] \
[--mapping-profile <profile>] [--confirm-unresolved <element>] \
[--property <key=value>] [--acknowledge-partial] [--wait]
| Argument | Description |
|---|---|
--point <point> | Point to restore. Required. |
--mode <mode> | Volume or instance build a new resource, bucket a new bucket, in_place overwrites the existing resource (default: volume) One of volume, instance, in_place, bucket. |
--name <name> | Name for the resource this creates; for mode bucket, the new bucket, which must not already exist. |
--volume-type <type> | Volume type for restored volumes. |
--availability-zone <az> | Target availability zone. |
--tier <tier> | Restore from this tier, chunk or rbd, instead of the fastest available source. One of chunk, rbd. |
--flavor <flavor> | Restore as new on a different flavor (mode instance) |
--networks <network,...> | Restore as new onto these networks: one for every port, or one per port in order (mode instance; comma-separated, repeatable) |
--key-name <key-pair> | Restore as new with a different key pair (mode instance) |
--security-groups <group,...> | Restore as new behind these security groups, by name or id (mode instance; comma-separated, repeatable) |
--disks <resource-id,...> | Restore only these disks of the point, by resource id (mode volume or instance; comma-separated, repeatable) |
--target-project <project> | Restore into another project (operator; needs a system-scoped token) |
--mapping-profile <profile> | Recovery profile, by name, that translates flavors, networks and volume types for a point another cloud wrote. |
--confirm-unresolved <element> | Accept an element the recovery profile could not map, as element or element:source (repeatable) |
--property <key=value> | Metadata on the created resource (repeatable) |
--acknowledge-partial | Proceed even though the point does not cover every resource it was meant to. |
--wait | Wait for the restore to finish. |
restore estimate
Estimate how long a restore would take, without starting one.
openstack backup restore estimate --point <point>
| Argument | Description |
|---|---|
--point <point> | Point to estimate. Required. |
restore list
List restores.
openstack backup restore list [--point <point>] [--state <state>] \
[--limit <limit>] [--project <project>]
| Argument | Description |
|---|---|
--point <point> | Only restores of this point. |
--state <state> | |
--limit <limit> | |
--project <project> | Another project's restores (operator; needs a system-scoped token) |
restore show
Show a restore, including anything it created.
openstack backup restore show <restore>
| Argument | Description |
|---|---|
<restore> | Restore id. |