This command sets the fs as not joinable and fails all ranks. This is a simpler
command than the typical sequence: (a) set fs not joinable; (b) iterate through
and fail ranks. It also does this in a single FSMap update.
Fixes: http://tracker.ceph.com/issues/37085
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
The dashboard ganesha management code requires the rados:// config
URL for the nfs-ganesha daemons. Ensure that we populate that.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
rook is lowercasing this tag in the CRD in response to review comments.
Change the orchestrator accordingly.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Changes needed to replace the obsolete get_config by the new get_module_option manager module method
Signed-off-by: Juan Miguel Olmo Martínez <jolmomar@redhat.com>
If we search for myaddrs, then we will detect our rank properly if we were
[a,b] and the a addr got removed.
Signed-off-by: Sage Weil <sage@redhat.com>
If *peers* are communicating, i.e. there may be bidirectional connection
attempts, we must use the same protocol version from both ends or else
we will get very confused.
Fix this by forcing the use of v1 when we
- are bound to a v1 endpoint only (people can't connect to us via v2)
- we are connecting to a *peer*
If it is a non-peer, then connections are uni-directional. If we both
have v2, we will both use v2.
If we ever switch to [v2,v1], it will be as part of a restart.
Signed-off-by: Sage Weil <sage@redhat.com>
If the ms_bind_msgr2 option is enabled, and all mons are nautilus,
raise a health alert if any mons aren't bound to msgr2 addresses.
Whitelist tests that mon_bind_addrvec=false or mon_bind_msgr2=false.
Signed-off-by: Sage Weil <sage@redhat.com>
This switches any monitor listening on v1:...:6789 only to v2 (on
3300) + v1 (on 6789). It should be run after upgrading all monitors
to nautilus.
Signed-off-by: Sage Weil <sage@redhat.com>
This is to prevent fmtflags changes from a previous use (e.g. debug message)
leaking into the the next (i.e. the next debug message).
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
mgr/dashboard: Filter out tasks depending on permissions
Reviewed-by: Ricardo Dias <rdias@suse.com>
Reviewed-by: Sebastian Wagner <swagner@suse.com>
Reviewed-by: Tatjana Dehler <tdehler@suse.com>
If we remove the last item of bucket, there should still be
one final entry in the __weights__ field of __weight_set__.
Free the corresponding memory before we __null__ the pointer.
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
This is to facilitate service start/stop/reload (for example,
after altering NFS Ganesha configuration, we need to trigger
a service reload).
Signed-off-by: Tim Serong <tserong@suse.com>
All the write/flush in SocketConnection should be serialized, or nasty
segment-fault could happen in seastar which is hard to diagnose.
This fix serializes the writes in `handle_keepalive2()` with other
writes in the open state.
Signed-off-by: Yingxin Cheng <yingxincheng@gmail.com>
pick_address.cc reference cct->_conf, so it should include the
definition of of CephContext explicitly instead of relying on the fact
that some headers it includes include its definition.
Signed-off-by: Kefu Chai <kchai@redhat.com>
in pick_address.cc, set_val_or_die() is called to set the "public_addr"
and/or "cluster_addr". but since the sharded ConfigProxy cannot
set_val_or_die() synchronously, we cannot do it this way.
we need to fix pick_addresses() so it works with seastar later. either
by making the involved function return future<>, or making the set_val()
call non-blocking or synchronous if we can go with 1:1 mapping
threading model.
but at this moment, let's just disable this feature in crimson-osd.
Signed-off-by: Kefu Chai <kchai@redhat.com>
unittest_seastar_messenger is failed since updated the seastar repo.
std::system_error (error system:22, Invalid argument) will be raised
from `listener->accept()`, which will be re-raised and not be captured.
This fix is to print error logs for the unexpected exceptions in
`SocketMessenger::start()` and not raise them again.
Signed-off-by: Yingxin Cheng <yingxincheng@gmail.com>
socket_addr is *our* socket address; target_addr is the remote peer's
socket address. We are inferring the remote's IP address here.
Fixes: http://tracker.ceph.com/issues/37882
Signed-off-by: Sage Weil <sage@redhat.com>
* refs/pull/25879/head:
client: define template to declare ref
client: use message smart ptr instead of raw ptr
client: make map lookup optimizations
client: do not clear message payload
client: clean up variable name
*: use std::string_view instead of char ptr
Reviewed-by: Sage Weil <sage@redhat.com>
