This isn't run automatically by anything yet. Note that it's also a best-effort
thing; passing doesn't guarantee there are tests. It can be pretty easily fooled
if the command is a common word which shows up in specifying other things,
for instance.
Signed-off-by: Greg Farnum <gfarnum@redhat.com>
Most of this is independent of the PGMonitor, so move it into
PGMap and strip out those few bits. It'll come in handy shortly
when I move "ceph osd df" into the mgr.
Signed-off-by: Greg Farnum <gfarnum@redhat.com>
Blob id exceeds maximum value, this patch is to prevent it.
Fixes: http://tracker.ceph.com/issues/19555
Signed-off-by: Xiaoyan Li <xiaoyan.li@intel.com>
osd: heartbeat with packets large enough to require working jumbo frames
Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Haomai Wang <haomai@xsky.com>
Reviewed-by: Josh Durgin <jdurgin@redhat.com>
msg/async: fix deleted_conn is out of sync with conns
Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Alexey Sheplyakov <asheplyakov@mirantis.com>
We already have a public method `set` for setting values in the RGW env
map, making the map as a read only. In addition:
- Added const to other methods in `RGWEnv` which are getters
- `req_info::init_meta` also reused the same `iter` name for an internal
find, changed the var name
- `add_grants_headers` now accepts an RGWEnv instead of the map
- use range based for loops wherever the code was changed
- req_info holds a ptr to a const RGWEnv, since we seem to only read the
values from it
Signed-off-by: Abhishek Lekshmanan <abhishek@suse.com>
The SUSE and Fedora packaging guidelines specify that subpackages _may_ have
their own License line if it is necessary. (Hopefully it's not, because
maintaining one License line is much easier.)
Signed-off-by: Nathan Cutler <ncutler@suse.com>
Although "which" is only used in the tests, we need the dependency
to be picked up by install-deps.sh.
This is a stopgap measure until we can get rid of the dependency entirely by
replacing "which" with "type -P" in the tests.
Fixes: http://tracker.ceph.com/issues/20127
Signed-off-by: Nathan Cutler <ncutler@suse.com>
Returning the number of items in the vector is not required and can be
calculated since the caller has access to the vector.
Signed-off-by: Brad Hubbard <bhubbard@redhat.com>
We get periodic reports that users somehow misconfigure one of their switches
so that it drops jumbo frames, yet the servers are still passing them along. In
that case, MOSDOp messages generally don't get through because they are much
larger than the 1500-byte non-jumbo limit, but the MOSDPing messages have kept
going (as they are very small and dispatched independently, even when the
server is willing to make jumbo frames). This means peer OSDs won't mark down
the ones behind the broken switch, despite all IO hanging.
Push the MOSDPing message size over the 1500-byte limit so that anybody in
this scenario will see the OSDs stuck behind a bad switch get marked down.
Fixes: http://tracker.ceph.com/issues/20087
Signed-off-by: Greg Farnum <gfarnum@redhat.com>
