----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/37512/#review95591 -----------------------------------------------------------
This looks very thorough. Just some minor comments/nits around typos. docs/scheduler_http_api.md (line 12) <https://reviews.apache.org/r/37512/#comment150624> s/via one/via docs/scheduler_http_api.md (line 14) <https://reviews.apache.org/r/37512/#comment150632> Since we refer to this endpoint via name "/scheduler" in the rest of the document. Why not add this line at the end after the fully qualified URL : We would be referring to this endpoint by its suffix "/scheduler" in the rest of the document. docs/scheduler_http_api.md (line 16) <https://reviews.apache.org/r/37512/#comment150626> How about this : This endpoint accepts HTTP POST requests with data encoded as JSON (Content-Type: application/json) or binary protobuf (Content-Type: application/x-protobuf). The first request that a scheduler sends to “/scheduler” endpoint is called SUBSCRIBE and results in a streaming response (Transfer-Encoding: chunked). **Schedulers are expected to keep the subscription connection open as long as possible (barring errors in network, software, hardware etc.) and incrementally process the response** (NOTE: HTTP client libraries that can only parse the response after the connection is closed cannot be used). For the encoding used, please refer to “Events” section below. docs/scheduler_http_api.md (line 19) <https://reviews.apache.org/r/37512/#comment150627> s/In line with the HTTP standard, a "202 Accepted"/An "Accepted" s/response just/response docs/scheduler_http_api.md (line 24) <https://reviews.apache.org/r/37512/#comment150629> s/exact protobuf/protobuf docs/scheduler_http_api.md (line 28) <https://reviews.apache.org/r/37512/#comment150631> s/very first/first docs/scheduler_http_api.md (line 30) <https://reviews.apache.org/r/37512/#comment150633> How about this ? To subscribe with the master, the scheduler sends a HTTP POST request with encoded `SUBSCRIBE` message containing the required field FrameworkInfo. Note that if "subscribe.framework_info.id" is not set, Master considers the scheduler as a new one and subscribes it by assigning a FrameworkID. The HTTP response is a stream with RecordIO encoding, with the first event being `SUBSCRIBED` event (see details in **Events** section). docs/scheduler_http_api.md (line 72) <https://reviews.apache.org/r/37512/#comment150634> s/with the same FrameworkID/containing the same docs/scheduler_http_api.md (line 76) <https://reviews.apache.org/r/37512/#comment150635> s/a HTTP 4xx response/, a HTTP 4xx response docs/scheduler_http_api.md (line 78) <https://reviews.apache.org/r/37512/#comment150636> s/result in/result in a s/"403 Forbidden" response instead of "202 Accepted" response/"403 Forbidden" instead of "202 Accepted" response s/“400 Bad Request”/"400 Bad Request" response s/bad/missing docs/scheduler_http_api.md (line 81) <https://reviews.apache.org/r/37512/#comment150649> How about: When Mesos receives this request, it will shut down all executors (and consequently kill tasks) and remove persistent volumes (if requested). The framework is then removed and all open connections from this scheduler to the Master are closed. docs/scheduler_http_api.md (line 97) <https://reviews.apache.org/r/37512/#comment150653> The other example don't have this line. Can we add this to the other example call except Subscribe too ? docs/scheduler_http_api.md (line 156) <https://reviews.apache.org/r/37512/#comment150656> How about : Sent by the scheduler to remove any/all filters that it has previously set via `ACCEPT` or `DECLINE` calls. docs/scheduler_http_api.md (line 224) <https://reviews.apache.org/r/37512/#comment150661> Kill "Note:uuid is needed to send an acknowledgement" as it is redundant. docs/scheduler_http_api.md (line 249) <https://reviews.apache.org/r/37512/#comment150662> s/status of a list of/status of docs/scheduler_http_api.md (line 333) <https://reviews.apache.org/r/37512/#comment150664> s/Very first event/The first event docs/scheduler_http_api.md (line 373) <https://reviews.apache.org/r/37512/#comment150665> Sent by the master whenever there is a status update that is generated by the executor, slave or master. docs/scheduler_http_api.md (line 414) <https://reviews.apache.org/r/37512/#comment150671> Note that there is no guaranteed order between the `FAILURE`, `UPDATE` and `RESCIND` events. docs/scheduler_http_api.md (line 444) <https://reviews.apache.org/r/37512/#comment150675> s/working/alive s/This will also help/This also helps docs/scheduler_http_api.md (line 472) <https://reviews.apache.org/r/37512/#comment150678> s/for any HTTP/to any HTTP docs/scheduler_http_api.md (line 521) <https://reviews.apache.org/r/37512/#comment150680> s/of a cluster/for a cluster - Anand Mazumdar On Aug. 17, 2015, 6:26 p.m., Vinod Kone wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/37512/ > ----------------------------------------------------------- > > (Updated Aug. 17, 2015, 6:26 p.m.) > > > Review request for mesos, Anand Mazumdar and Ben Mahler. > > > Bugs: MESOS-3281 > https://issues.apache.org/jira/browse/MESOS-3281 > > > Repository: mesos > > > Description > ------- > > Mostly copy pasted from the design doc with some formatting. > > > Diffs > ----- > > docs/scheduler_http_api.md PRE-CREATION > > Diff: https://reviews.apache.org/r/37512/diff/ > > > Testing > ------- > > gist: https://gist.github.com/vinodkone/5ae803ee35274ca46a3e > > > Thanks, > > Vinod Kone > >