CP-46179 backport for deterministic backup VDI #5552
Conversation
|
|
||
| let namespace = | ||
| let ns = "e93e0639-2bdb-4a59-8b46-352b3f408c19" in | ||
| Uuidm.(of_string ns |> Option.get) |
There was a problem hiding this comment.
The Uuidx library is available as Uuid in yangtze. Patching it seems more involved than doing it here, so I think this change makes more sense than backporting it to https://github.com/xapi-project/xen-api-libs-transitional/blob/2.25-lcm/uuid/uuid.ml
| @@ -48,6 +48,29 @@ function usage { | |||
| exit 1 | |||
| } | |||
|
|
|||
| function uuid5 { | |||
There was a problem hiding this comment.
missing hash of original commit in the commit message: 084c6a9
| @@ -74,7 +74,7 @@ function validate_vdi_uuid { | |||
| function test_sr { | |||
| sr_uuid_found=$(${XE} sr-list uuid="$1" --minimal) | |||
| if [ "${sr_uuid_found}" != "$1" ]; then | |||
| echo Invalid SR UUID specified: $1 | |||
| echo "Invalid SR UUID specified: $1" | |||
There was a problem hiding this comment.
missing hash of original commit in the commit message: f88fa12
| @@ -39,6 +39,7 @@ function usage { | |||
| echo " -k: Number of older backups to preserve (default: ${history_kept})" | |||
| echo " -n: Just try to find a backup VDI and stop the script after that" | |||
| echo " -f Force backup even when less than 10% free capacity is left on the backup VDI" | |||
| echo " -y: Assume non-interactive mode and yes to all questions" | |||
There was a problem hiding this comment.
missing hash of original commit in the commit message: 4fe8ffa
|
Backports look faithful to me, I've only got the small nit of the missing original hashes in the commit messages. Can this be merged as is, or does it depend on some xsconsole changes? |
|
This can be merged but should not be released as a hot fix because it might depend on xsconsole changes. Which creates the additional problem of creating an ordering problem. But I think we should at least get the back port into our repo somehow. |
Backport of bbc5a1f When creating a backup VDI on an SR we want to derive the VDI's UUID from the SR such that we know how to find it later on the SR. Implement a hash that derives the VDI UUID and pass this UUID to SM to be used. Currently the creation of this VDI in Xapi is detected based on its name-label: "Pool Metadata Backup". The SM stack usually creates a new random UUID but will use a given UUID passed as "vdi_uuid" for its "vdi_create" command. The difficulty in the implementation is that "vdi_uuid" is used in other commands as well and we have to make sure to set it only in the intended context. It is not straight forward to pass the UUID from Xapi_vdi.create down to Sm. Signed-off-by: Christian Lindig <[email protected]>
lindig
force-pushed
the
private/christianlin/1.249-lcm
branch
from
7ab7414 to
7d409fd
Compare
|
I've updated the commit messages to include the hashes of the original commits. |
Backport of 65f152d We introduce deterministic UUID for backup VDIs to avoid searching for VDIs. Add this capability to the scripts that create and restore backups. In case the VDI with the expected UUID does not exist, let the user confirm probing for it. Signed-off-by: Christian Lindig <[email protected]>
Backport of c2f7456 Signed-off-by: Christian Lindig <[email protected]>
Backport of 7f1d315 We have added an interactive prompt in case the backup VDI was created by an earlier version and does not match the expected UUID. To facilitate scripting, add a -y (yes) flag that assumes a confirmation. Signed-off-by: Christian Lindig <[email protected]>
lindig
force-pushed
the
private/christianlin/1.249-lcm
branch
from
7d409fd to
f0eb272
Compare
| @@ -113,7 +113,23 @@ let make_call ?driver_params ?sr_sm_config ?vdi_sm_config ?vdi_type | |||
| may (fun self -> Db.VDI.get_location ~__context ~self) vdi_ref | |||
| in | |||
| let vdi_uuid = | |||
| may (fun self -> Db.VDI.get_uuid ~__context ~self) vdi_ref | |||
| let __FUNCTION__ = "Sm_exec.make_call" in | |||
There was a problem hiding this comment.
why redefining __FUNCTION__?
There was a problem hiding this comment.
It is not defined in the compiler version on Yangtze. So now I don't have to change the code in the back port.
| @@ -688,16 +688,31 @@ module SMAPIv1 = struct | |||
| vdi_info_from_db ~__context (Db.VDI.get_by_uuid ~__context ~uuid) | |||
|
|
|||
There was a problem hiding this comment.
this storage_access.ml file corresponds to storage_smapiv1.ml in master, no smapiv3 back then?
We want to use a predictable VDI for metadata backup. This is the backport of that feature to Yangtze.