github.com/golang-jwt/jwt adds go mod support startig from 4.0.0,
and it's backwards-compatible with existing v3.x.y tags.
Signed-off-by: Benjamin Wang <wachao@vmware.com>
A WAL object was closed by defer, however the WAL was rewritten afterwards,
so defer closed already closed WAL but not the new one. It caused a data
race between writing file and cleaning up a temporary test directory,
which led to a non-deterministic bug.
Fixes#14332
Signed-off-by: Vladimir Sokolov <vsvastey@gmail.com>
The alarm list is the only exception that doesn't move consistent_index
forward. The reproduction steps are as simple as,
```
etcd --snapshot-count=5 &
for i in {1..6}; do etcdctl alarm list; done
kill -9 <etcd_pid>
etcd
```
Signed-off-by: Benjamin Wang <wachao@vmware.com>
For a cluster with only one member, the raft always send identical
unstable entries and committed entries to etcdserver, and etcd
responds to the client once it finishes (actually partially) the
applying workflow.
When the client receives the response, it doesn't mean etcd has already
successfully saved the data, including BoltDB and WAL, because:
1. etcd commits the boltDB transaction periodically instead of on each request;
2. etcd saves WAL entries in parallel with applying the committed entries.
Accordingly, it may run into a situation of data loss when the etcd crashes
immediately after responding to the client and before the boltDB and WAL
successfully save the data to disk.
Note that this issue can only happen for clusters with only one member.
For clusters with multiple members, it isn't an issue, because etcd will
not commit & apply the data before it being replicated to majority members.
When the client receives the response, it means the data must have been applied.
It further means the data must have been committed.
Note: for clusters with multiple members, the raft will never send identical
unstable entries and committed entries to etcdserver.
Signed-off-by: Benjamin Wang <wachao@vmware.com>
When we can't reach quorum, we were waiting forever and never sending
the systemd notify message. As a result, systemd would eventually time out
and restart the etcd process which likely would make the unhealthy cluster
in an even worse state
Improves #13785
Signed-off-by: Nicolai Moore <niconorsk@gmail.com>
Only `net.TCPConn` supports `SetKeepAlive` and `SetKeepAlivePeriod`
by default, so if you want to warp multiple layers of net.Listener,
the `keepaliveListener` should be the one which is closest to the
original `net.Listener` implementation, namely `TCPListener`.
Signed-off-by: Benjamin Wang <wachao@vmware.com>
As protobuf doesn't have required field, user may send an empty
WatchRequest by mistake. Currently, etcd will ignore the invalid request
and keep the stream opening. If we don't reject the invalid request by
closing the stream, it would be better to leave a log there.
This commit also fixes a typo in the comment.
Signed-off-by: spacewander <spacewanderlzx@gmail.com>
inspired by https://github.com/grpc-ecosystem/go-grpc-middleware
Signed-off-by: stefanbo <stefan_bo@163.com>
Update flag name
1. add changelog
2. update flag name to `experimental-enable-grpc-debug`
Signed-off-by: stefan bo <stefan_bo@163.com>
Update CHANGELOG-3.6.md
Signed-off-by: stefan bo <stefan_bo@163.com>
change flag name
Signed-off-by: stefan bo <stefan_bo@163.com>
Problem: We pass grpc context down to applier in readonly serializable txn.
This context can be cancelled for example due to timeout.
This will trigger panic inside applyTxn
Solution: Only panic for transactions with write operations
fixes https://github.com/etcd-io/etcd/issues/14110
Signed-off-by: Bogdan Kanivets <bkanivets@apple.com>
