Progress report for restore about longhorn HOT 8 CLOSED

@smallteeths backend merged.

from longhorn.

@shreesha21 Can you document the API reference for the restoration status for UI here?

from longhorn.

@smallteeths for now you should able to refer to https://github.com/longhorn/longhorn-manager/blob/master///api/model.go#L584

Each replica has it's own restoration status, and the volume can show the average.
Progress is in % as backup status.
State can be incomplete, complete, error. When it's incomplete, progress shows the detail percentage of completeness. When it's error, the error field should show the reason.

from longhorn.

Request: curl http://172.104.18.252:30766/v1/volumes
Response: The response contains all the fields of volume object along with restoreStatus as below:

   "data" : [
      {
         "restoreStatus" : [
            {
               "lastRestored" : "",
               "progress" : 0,
               "actions" : null,
               "replica" : "tcp://10.42.0.75:9502",
               "isRestoring" : false,
               "error" : "",
               "state" : "",
               "links" : null,
               "filename" : ""
            },
            {
               "replica" : "tcp://10.42.2.132:9502",
               "isRestoring" : true,
               "actions" : null,
               "progress" : 73,
               "lastRestored" : "",
               "filename" : "volume-snap-76cf801b-279e-4a0b-a36b-8c78cb279135.img",
               "links" : null,
               "error" : "",
               "state" : "incomplete"
            }
            ...
            ...
         ],
}

from longhorn.

@shreesha21 We need to update the replica field in the manager API to report the replica name instead of IP, otherwise it's hard for UI to figure out which is which. Only need to update the manager API, no change needed for the engine.

from longhorn.

@smallteeths longhorn/longhorn-manager#355 has been merged, now API shows replica name instead of IP.

from longhorn.

@smallteeths reports that the replica states during the restoration are switching between RW and ``(empty). We should report empty since it's not ready yet.

@shreesha21 ^^

from longhorn.

Validattion: PASSED

Steps to test:

• Create a volume, attach it to a node, format it, mount it to a directory on attached node, write some data to it.
• In Settings page, make sure backup target is correctly configured.
• In Volume Details page, Click Create Backup, and wait for backup to complete.
• In Backup page, click on volume name
• From the right drop down menu of the backup, click Restore.
• Enter Name, and set Number of replicas, and click OK

Expected result: Restoration should start, and restoration progress should be visible in
1 - Volume Details page at Volume Details section
2 - Volume Details page at Replicas section for every volume replica.
3 - Volume page under volume status.

from longhorn.