Difference in load configuration for watch delay tests show how huge the
impact is. Even with random write scheduler grpc under http
server can only handle 500 KB with 2 seconds delay. On the other hand,
separate grpc server easily hits 10, 100 or even 1000 MB within 100 miliseconds.
Priority write scheduler that was used in most previous releases
is far worse than random one.
Tests configured to only 5 MB to avoid flakes and taking too long to fill
etcd.
Signed-off-by: Marek Siarkowicz <siarkowicz@google.com>
There are two goroutines accessing the `gs` grpc server var. Before
insecure `gs` server start, the `gs` can be changed to secure server and
then the client will fail to connect to etcd with insecure request. It
is data-race. We should use argument for reference in the new goroutine.
fix: #15495
Signed-off-by: Wei Fu <fuweid89@gmail.com>
(cherry picked from commit a9988e2625eede1af81d189b5f2ecf7d4af3edf1)
Signed-off-by: Wei Fu <fuweid89@gmail.com>
To avoid inconsistant behavior during cluster upgrade we are feature
gating persistance behind cluster version. This should ensure that
all cluster members are upgraded to v3.6 before changing behavior.
To allow backporting this fix to v3.5 we are also introducing flag
--experimental-enable-lease-checkpoint-persist that will allow for
smooth upgrade in v3.5 clusters with this feature enabled.
Signed-off-by: Marek Siarkowicz <siarkowicz@google.com>
Ensure the client which access etcd via grpc-gateway won't
be limited by the MaxCallRecvMsgSize. Here we choose the same
default value of etcdcli as grpc-gateway's MaxCallRecvMsgSize.
Fix https://github.com/etcd-io/etcd/issues/12576
This fixes etcd being unable to send any message longer than 64 KB as
a notification over the websocket. This was because the older version
of grpc-websocket-proxy was used and WithMaxRespBodyBufferSize option
wasn't set.
This makes it possible to run an etcd node for testing and development
without placing lots of load on the file system.
Fixes#11930.
Signed-off-by: David Crawshaw <crawshaw@tailscale.com>
Make log level configurable, and deprecate "debug" flag in v3.5.
And adds more warnings on flags that's being deprecated in v3.5.
Signed-off-by: Gyuho Lee <leegyuho@amazon.com>
This fixes an index out-of-bounds panic caused when using the embed
package and the zap logger. When a TLS handshake error is logged, the
slice for cert ip addresses is allocated with capacity but no length, so
subsequent index access causes the panic, and doesn't surface the TLS
handshake error to the user.
Fixes#10932
To remove the dependency on ghodss/yaml. Replaced this dependency with sigs.k8s.io/yaml.
This wil help to remove the ghodss/yaml dependency from main kubernetes repository.
xref: https://github.com/kubernetes/kubernetes/issues/77024
This commit lets grpc gateway return a correct error to clients.
Even if a client has a cert with non empty CN, current gateway returns
an error like below:
```
$ curl --cacert ./integration/fixtures/ca.crt --cert ./integration/fixtures/server.crt --key ./integration/fixtures/server.key.insecure https://localhost:2379/v3/kv/put -X POST -d '{"key": "fromcurl", "value": "test"}'
{"error":"etcdserver: user name is empty","code":3}
```
This is because etcd ignores CN from gateway connection.
The error will be like this:
```
$ curl --cacert ./integration/fixtures/ca.crt --cert ./integration/fixtures/server.crt --key ./integration/fixtures/server.key.insecure https://localhost:2379/v3/kv/put -X POST -d '{"key": "fromcurl", "value": "test"}'
CommonName of client sending a request against gateway will be ignored and not used as expected
```
The error will be returned if the server is enabling auth and gRPC
gateway.
