Update patch set 4
Patch Set 4: (3 comments) Patch-set: 4
This commit is contained in:
parent
24365c0829
commit
93468ce819
|
@ -0,0 +1,73 @@
|
|||
{
|
||||
"comments": [
|
||||
{
|
||||
"key": {
|
||||
"uuid": "9fb8cfa7_38555e51",
|
||||
"filename": "specs/train/secret-consumers.rst",
|
||||
"patchSetId": 4
|
||||
},
|
||||
"lineNbr": 103,
|
||||
"author": {
|
||||
"id": 7973
|
||||
},
|
||||
"writtenOn": "2019-06-03T16:34:53Z",
|
||||
"side": 1,
|
||||
"message": "Adding this per @Ade Lee\u0027s comments on Patch 3, however I\u0027m not sure if Barbican should consider the tuple (serive, resource_type, resource_id) to identify a unique consumer or just the resource_id.\n\nAt the PTG we talked about assuming that resource_id would be unique in a cloud since all projects use UUIDs, so the chances of two services having the same resource_id for different resources would be unlikely.",
|
||||
"range": {
|
||||
"startLine": 102,
|
||||
"startChar": 0,
|
||||
"endLine": 103,
|
||||
"endChar": 41
|
||||
},
|
||||
"revId": "4b5c70a0c630e410700bbf509b68176911ec86d6",
|
||||
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543",
|
||||
"unresolved": false
|
||||
},
|
||||
{
|
||||
"key": {
|
||||
"uuid": "9fb8cfa7_db9f5cb2",
|
||||
"filename": "specs/train/secret-consumers.rst",
|
||||
"patchSetId": 4
|
||||
},
|
||||
"lineNbr": 109,
|
||||
"author": {
|
||||
"id": 7973
|
||||
},
|
||||
"writtenOn": "2019-06-03T16:34:53Z",
|
||||
"side": 1,
|
||||
"message": "This will be tricky. The advantage of using these values instead of just a URL is that we should, in theory, be able to recreate a URL using Version Discovery, so it would be future-proof. I\u0027m not sure if all OpenStack APIs are written to be navigated with HATEOS links though?\n\nOn the other hand, instead of using a `resource_type` we may be able to use a `resource_path` instead. For example:\n\nIn glance, the Keystone Catalog would point to `/` which returns the current Version as ending in `v2`. The path to get an image is `v2/images/{image_id}` so resource_type\u003dimages would work to get the path from version discovery -\u003e entity URL.\n\nIn octavia, however, the Keystone Catalog would point to `/` which returns the current version as ending in `v2`, but the pat to get a load balancer is `v2/lbass/loadbalancers/{loadbalancer_id}. The response for `v2` is not documented so it\u0027s not clear if there is a way to map the type `loadbalancers` to the path `lbass/loadbalancers`. In this case the resource_type\u003dloadbalancers would not work so they would use resource_path\u003d\"lbass/loadbalancers\" instead. Thoughts?",
|
||||
"range": {
|
||||
"startLine": 103,
|
||||
"startChar": 43,
|
||||
"endLine": 109,
|
||||
"endChar": 32
|
||||
},
|
||||
"revId": "4b5c70a0c630e410700bbf509b68176911ec86d6",
|
||||
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543",
|
||||
"unresolved": false
|
||||
},
|
||||
{
|
||||
"key": {
|
||||
"uuid": "9fb8cfa7_d82f62f4",
|
||||
"filename": "specs/train/secret-consumers.rst",
|
||||
"patchSetId": 4
|
||||
},
|
||||
"lineNbr": 209,
|
||||
"author": {
|
||||
"id": 7973
|
||||
},
|
||||
"writtenOn": "2019-06-03T16:34:53Z",
|
||||
"side": 1,
|
||||
"message": "If we do want to consider all 3 values together as an identifier (see comment above), then maybe it makes sense to use all 3 parts in the DELETE path?",
|
||||
"range": {
|
||||
"startLine": 209,
|
||||
"startChar": 45,
|
||||
"endLine": 209,
|
||||
"endChar": 84
|
||||
},
|
||||
"revId": "4b5c70a0c630e410700bbf509b68176911ec86d6",
|
||||
"serverId": "4a232e18-c5a9-48ee-94c0-e04e7cca6543",
|
||||
"unresolved": false
|
||||
}
|
||||
]
|
||||
}
|
Loading…
Reference in New Issue