Support the following settings:
* mode: both, client mode, server mode;
* set server address;
* rounds of messages;
* keepalive-ratio;
* block size;
* io-depth;
Signed-off-by: Yingxin Cheng <yingxincheng@gmail.com>
"float:none" is changing the order of the table buttons.
By reverting this change and setting .oadatatableactions to "float:left",
we get the same result as intended before.
Signed-off-by: Tiago Melo <tmelo@suse.com>
python-jinja2 dependency was added for the dashboard v1 in 87399bea. but
dashboard v2, which is now known as "dashboard" has taken the place of
the old dashboard. and dashboard now does not use jinja2, so we should
drop this dependency.
Signed-off-by: Kefu Chai <kchai@redhat.com>
The new unittest_seastar_messenger will print out the count of keepalive
attempts at the end of the execution.
Signed-off-by: Yingxin Cheng <yingxincheng@gmail.com>
Implement the sharded crimson-messenger:
* Sharded Messenger: provides shared-nothing Messenger for each shard,
it's interfaces are symmetric to be called, any modifications will be
applied to all shards.
* Sharded/non-sharded Dispatcher interface: allow connections to be
dispatched, and related resources (such as Session) to be managed in
its own shard or not.
* Sharded Connection: A connection only lives at one dedicated core
during its lifecycle. It's sharded by its peer_IP in this PoC, because
peer port and nonce are not available when a socket is accepted. While
its interfaces are safe to be called from all shards.
* Replace `boost::intrusive_ptr` by seastar native smart ptrs for
`Connection` and `SocketConnection`, because they need to be
destructed from its original core.
* Unit test: establish multiple connections on both client and server
sides, they runs concurrently and creates sessions that are also
following shared-nothing design.
Signed-off-by: Yingxin Cheng <yingxincheng@gmail.com>
change following plugins' Architecture to "all", as they are pure python
plugins, and are architecture-independent packages.
- ceph-mgr-diskprediction-cloud
- ceph-mgr-diskprediction-local
- ceph-mgr-rook
this also matches their couterparts in ceph.spec.in, where we have
```
BuildArch: noarch
```
Signed-off-by: Kefu Chai <kchai@redhat.com>
* refs/pull/26336/head:
qa/tasks/keystone.py: no need for notcmalloc in example
qa/suites/rgw/tempest/tasks/rgw_tempest: no need for notcmalloc
Reviewed-by: Alfredo Deza <adeza@redhat.com>
Reviewed-by: Yehuda Sadeh <yehuda@redhat.com>
The same scenario that was possible on client_ident is also possible
for reconnect: on the accepting side we recognized the client_addrs
but the client is trying to reach someone else. Since the cookie is a
random 64-bit value it is unlikely, but still possible, that we will
match an existing session that is not for the client.
Signed-off-by: Sage Weil <sage@redhat.com>
If we get a client_ident frame, and they are trying to talk to someone
else, drop the connection.
This is an inelegant workaround to http://tracker.ceph.com/issues/38247.
A nicer fix would be to restructure the protocol so that the client knows
who they connected to before they try to open a session. That is a bigger
change that can follow...
Fixes: http://tracker.ceph.com/issues/38247
Signed-off-by: Sage Weil <sage@redhat.com>
The true parent of a migrating parent is actually the parent of the
migration source image. There are other API methods available to
retreive the details of the migration source.
Fixes: http://tracker.ceph.com/issues/37998
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
