the python bindings are compiled using python3-cython, and packaged as
python3-{rados,rbd,cephfs,rgw}, so there is no need and is wrong to
claim that they are python2 compatible.
Signed-off-by: Kefu Chai <kchai@redhat.com>
clang++ complains with
```
crimson/os/seastore/root_block.h💯3: warning: control reaches end of non-void function [-Wreturn-type]
}
^
```
Signed-off-by: Kefu Chai <kchai@redhat.com>
The specific_result test hardcodes a particular output. However,
the input buffer used by the test is currently initialized to a
different byte string depending host the host byte order, which
makes the expected result not match on big-endian hosts.
Fixed by always initializing the input buffer to the same string.
Signed-off-by: Ulrich Weigand <ulrich.weigand@de.ibm.com>
When decoding an entity_addrvec_t with marker 1, we just have
a single (non-legacy) entity_addr_t. This should be decoded
exactly the same as done by entity_addr_t::decode, but it
currently is not. Specifically, the sa_family member of
the sockaddr is not converted from the on-wire little-endian
format to host byte order (as done by entity_addr_t::decode).
Fixed by using the same code as in entity_addr_t::decode.
Signed-off-by: Ulrich Weigand <ulrich.weigand@de.ibm.com>
Introduced ZonedFreelistManager, which tracks the write pointer and the number
of dead bytes per zone.
Signed-off-by: Abutalib Aghayev <agayev@cs.cmu.edu>
in release build, `found` is not used at all. but GCC complains with a
`-Wunused-but-set-variable` warning.
Signed-off-by: Kefu Chai <kchai@redhat.com>
because we need to implement a tell command which forces osd to
send the latest pg stats to mgr, and the command returns the user with
the sequence id of the report, and `mgr::Client::report()` does not
return a future, so we have to update the seq id before sending
the report. the solution is to update the seq id in a separated
method, so in this change:
* add `const` to `WithStats::get_stats()
* add a dedicated method of `WithStats::update_stats()` to update
the stats to be collected by mgr::Client.
Signed-off-by: Kefu Chai <kchai@redhat.com>
Currently the list_network command and methods in cephadm only run and
parse ipv4 output from `ip route`.
This patch extends the list_network command and internal methods to be
ipv6 enabled. It now also checks `ip -6 route` and `ip -6 addr` to
gather gather all networks from both protocol families.
Signed-off-by: Matthew Oliver <moliver@suse.com>
crimson/os: pin the last cpu core for the alien worker threads
Reviewed-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
Reviewed-by: Deepika Upadhyay <dupadhya@redhat.com>
mgr/dashboard: Host delete action should be disabled if not managed by Orchestrator
Reviewed-by: Kiefer Chang <kiefer.chang@suse.com>
Reviewed-by: Stephan Müller <smueller@suse.com>
before this change, we assume that we have at least current_shared + 10
cores. but that's not always true. so in this change, the last core
is used for performing the alien tasks scheduled by reactor.
Signed-off-by: Kefu Chai <kchai@redhat.com>
This PR refactors the /monitoring page by introducing a new component which is responsible for rendering the tabs. All tabs have a dedicated route/URL now, no fragments are used anymore, thus they will act like the iSCSI and RBD pages.
Fixes: https://tracker.ceph.com/issues/45375
Signed-off-by: Volker Theile <vtheile@suse.com>
msg/async/ProtocolV2: take care of features when replacing the socket
Reviewed-by: Patrick Donnelly <pdonnell@redhat.com>
Reviewed-by: Neha Ojha <nojha@redhat.com>
Reviewed-by: Radoslaw Zarzynski <rzarzyns@redhat.com>
