CHANGELOG-3.4: add watch fragmentation

Signed-off-by: Gyuho Lee <gyuhox@gmail.com>
etcd-io · May 14, 2018 · 2a51072 · 2a51072
1 parent 56ec416
commit 2a51072
Showing 1 changed file with 16 additions and 1 deletion.
diff --git a/CHANGELOG-3.4.md b/CHANGELOG-3.4.md
@@ -199,8 +199,14 @@ See [security doc](https://github.com/coreos/etcd/blob/master/Documentation/op-g
 ### API
 
 - Add [`snapshot`](https://github.com/coreos/etcd/pull/9118) package for snapshot restore/save operations (see [`godoc.org/github.com/etcd/snapshot`](https://godoc.org/github.com/coreos/etcd/snapshot) for more).
-- Add [`watch_id` field to `etcdserverpb.WatchCreateRequest`](https://github.com/coreos/etcd/pull/9065), allow user-provided watch ID to `mvcc`.
+- Add [`watch_id` field to `etcdserverpb.WatchCreateRequest`](https://github.com/coreos/etcd/pull/9065) to allow user-provided watch ID to `mvcc`.
   - Corresponding `watch_id` is returned via `etcdserverpb.WatchResponse`, if any.
+- Add [`fragment` field to `etcdserverpb.WatchCreateRequest`](https://github.com/coreos/etcd/pull/9291) to request etcd server to [split watch events](https://github.com/coreos/etcd/issues/9294) when the total size of events exceeds `--max-request-bytes` flag value plus gRPC-overhead 512 bytes.
+  - The default server-side request bytes limit is `embed.DefaultMaxRequestBytes` which is 1.5 MiB plus gRPC-overhead 512 bytes.
+  - If watch response events exceed this server-side request limit and watch request is created with `fragment` field `true`, the server will split watch events into a set of chunks, each of which is a subset of watch events below server-side request limit.
+  - For example, watch response contains 10 events, where each event is 1 MiB. And server `--max-request-bytes` flag value is 1 MiB. Then, server will send 10 separate fragmented events to the client.
+  - For example, watch response contains 5 events, where each event is 2 MiB. And server `--max-request-bytes` flag value is 1 MiB and `clientv3.Config.MaxCallRecvMsgSize` is 1 MiB. Then, server will try to send 5 separate fragmented events to the client, and the client will error with `"code = ResourceExhausted desc = grpc: received message larger than max (...)"`.
+  - Client must implement fragmented watch event merge (which `clientv3` does in etcd v3.4).
 - Add [`raftAppliedIndex` field to `etcdserverpb.StatusResponse`](https://github.com/coreos/etcd/pull/9176) for current Raft applied index.
 - Add [`errors` field to `etcdserverpb.StatusResponse`](https://github.com/coreos/etcd/pull/9206) for server-side error.
   - e.g. `"etcdserver: no leader", "NOSPACE", "CORRUPT"`
@@ -224,6 +230,15 @@ See [security doc](https://github.com/coreos/etcd/blob/master/Documentation/op-g
 - Add [`CLUSTER_DEBUG` to enable test cluster logging](https://github.com/coreos/etcd/pull/9678).
   - Deprecated `capnslog` in integration tests.
 
+### client v3
+
+- Add [`WithFragment` `OpOption`](https://github.com/coreos/etcd/pull/9291) to support [watch events fragmentation](https://github.com/coreos/etcd/issues/9294) when the total size of events exceeds `--max-request-bytes` flag value plus gRPC-overhead 512 bytes.
+  - Watch fragmentation is disabled by default.
+  - The default server-side request bytes limit is `embed.DefaultMaxRequestBytes` which is 1.5 MiB plus gRPC-overhead 512 bytes.
+  - If watch response events exceed this server-side request limit and watch request is created with `fragment` field `true`, the server will split watch events into a set of chunks, each of which is a subset of watch events below server-side request limit.
+  - For example, watch response contains 10 events, where each event is 1 MiB. And server `--max-request-bytes` flag value is 1 MiB. Then, server will send 10 separate fragmented events to the client.
+  - For example, watch response contains 5 events, where each event is 2 MiB. And server `--max-request-bytes` flag value is 1 MiB and `clientv3.Config.MaxCallRecvMsgSize` is 1 MiB. Then, server will try to send 5 separate fragmented events to the client, and the client will error with `"code = ResourceExhausted desc = grpc: received message larger than max (...)"`.
+
 ### etcdctl v3
 
 - Add [`check datascale`](https://github.com/coreos/etcd/pull/9185) command.