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>
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.
Flags `--experimental-enable-v2v3` and '-enable-v2' will raise a warning in 3.5,
in 3.6 they are schedule for decomissioning, such that v2store can stop be written in 3.7.
Deprecation plan in: https://github.com/etcd-io/etcd/issues/12913
clientv3 logs (especially tests) were poluted with unattributed to testing.T log lines:
```
{"level":"warn","ts":"2021-04-29T12:42:11.055+0200","logger":"etcd-client","caller":"v3/retry_interceptor.go:64","msg":"retrying of unary invoker failed","target":"etcd-endpoints://0xc0000fafc0/#initially=[unix://localhost:m10]","attempt":0,"error":"rpc error: code = ResourceExhausted desc = etcdserver: mvcc: database space exceeded"}
```
The reasons were 2 fold:
- Interceptors were copying logger before "WithLogger" could modify it.
- We were not propagating the loggers in a few testing contexts.
So far each instance of embed server was overriding the grpc loggers and zap.global loggers.
It's counter intutitive that last created Embedded server was 'wining' and more-over it was breaking grpc expectation to change it "only" before the grpc stack is being used.
This PR introduces explicit call: `embed.Config::SetupGlobalLoggers()`, that changes the loggers where requested. The call is used by etcd main binary.
The immediate benefit from this change is reduction of test flakiness, as there were flakes due to not a proper logger being used across tests.
The flag protects etcd memory from being swapped out to disk.
This can happen in memory constrained systems where mmaped bbolt
area is natural condidate for swapping out.
This flag should provide better tail latency on the cost of higher RSS
ram usage. If the experiment is successful, the logic should get moved
into bbolt layer, where we can protect specific bbolt instances
(e.g. avoid protecting both during defragmentation).
In some environments, the CA is not able to sign certificates with both
'client auth' and 'server auth' extended usage parameters and so an operator
needs to be able to set a seperate client certificate to use when making
requests which is different to the certificate used for accepting requests.
This applies to both proxy and etcd member mode and is available as both a CLI
flag and config file field for peer TLS.
Signed-off-by: Ben Meier <ben.meier@oracle.com>
