Commit 3d7c69fb09 introduced a new OSDMap encoding/decoding scheme.
However, while the classic decoding function still kept building the
reverse name->pool map, the new decoding function did not, causing the
monitor to be unable to map pool names to pool ids.
This patch fixes this, by factoring out the loop responsible for
populating the 'name_pool' map, as well as calling 'calc_num_osds()', to
OSDMap::post_decode() and having this function called from both the
classic and the new decode functions.
Fixes: 7166
Signed-off-by: Joao Eduardo Luis <jecluis@gmail.com>
Automake puts ceph_common.sh into libdir/ceph, but the Red Hat packaging
was not capturing this file.
Add the libdir/ceph location to the RPM packaging.
Fixes#7117
Two tests were inadvertantly missing from the RPM packaging:
ceph_test_c_headers was added to ceph in
ffdd30efc8
ceph_test_get_blkdev_size was added to ceph in
324dd54534
Add both of these to the packaging.
Signed-off-by: Ken Dreyer <ken.dreyer@inktank.com>
Making sure a cluster supports primary_temp is complicated and we don't
have any of the machinery in place right now (nor a need to actually support
it). We don't have any mechanisms for setting it to begin with, so assert
that we never create anything with any such mapping in update_from_paxos()
to catch any errors.
Signed-off-by: Greg Farnum <greg@inktank.com>
This is not super-sophisticated, but it does basic mapping function
consistency checks and looks at the [pg|primary]_temp manipulations. If
we want to in the future, we can do these programmatically across a range
of pgids instead of just checking hash 0.
Signed-off-by: Greg Farnum <greg@inktank.com>
Switch _get_temp_osds to use pointers instead of references, and force callers
to check the out params instead of relying on a return code for if anything
was set (trying to use the return code when there are two possible outputs
does not provide useable semantics). For the new temp_primary out param, fill it
in from temp_primary if set, or from the pg_temp list if it's set, or leave
it blank if neither are.
Also, don't use pointers to heap elements. Just put the ints and vectors on
the stack, and assign/swap the out parameters with them. This is less
confusing and should be a bit faster in general.
Signed-off-by: Greg Farnum <greg@inktank.com>
Switch to use pointers for the out parameters instead of references.
These functions are still just pointing at the front of the generated
lists for the "primary" params, but now that all their callers respect
these outputs we can add programmatic leader assignment with just these
two functions.
Signed-off-by: Greg Farnum <greg@inktank.com>
And use pointers instead of references for out params.
Now pg_to_up_acting_osds and pg_to_acting_osds can plug in to this slightly
more real implementation, instead of making up their own. (We are still
just using the first member anyway, but we're about to plug it into
the bottom layer of functions.)
Signed-off-by: Greg Farnum <greg@inktank.com>
So that this works with future CRUSH changes, we copy the map and clear
out the primary_temp, then compare its output with the real map's output. If
they match, remove the primary_temp from the real map.
Signed-off-by: Greg Farnum <greg@inktank.com>
Bring our OSDMap encoding into the modern Ceph world! :) This is
fairly straightforward, but has a few rough edges:
Previously we had a "struct_v" which went at the beginning of the
OSDMap encoding, and then later on an ev "extended version" which
was used to store the more-frequently-changed OSDMap pieces. There
was no size information stored explicitly to let clients skip this,
but osd maps were always encoded into their own bufferlist before
being sent to clients, which had the same effect.
We now use the modern ENCODE_START three times:
1) for the overall OSDMap encoding,
2) for the client-usable portion of the map,
3) for the "extended" portion of the map
This will let us independently rev everything, which may come in
useful if we want to (for instance) add a "monitor" portion to the
map that the OSDs don't care about. It also makes adding new
client information a lot easier since older clients will still
be able to decode the map as a whole.
We may want to merge this OSDMAP_ENC feature with one of the others
we are creating during this cycle, since they're all very closely
related. That will also let us protect more naturally against old
clients getting a map they need to understand but can't (because
we only need the new map features-to-come when used with erasure-encoded
PGs, etc).
Signed-off-by: Greg Farnum <greg@inktank.com>
Use our pointer calling conventions instead of a reference for the
new version of the function.
Right now we're just setting the primaries equal to the first member
of up and acting (or -1 if none), but very shortly we'll modify our
private OSDMap functions to export them based on the contents of temp_primary.
While in general anybody querying for the mapping information will
need to pay attention to whom the primary is as well, we have lots
of callers who will need real code changes to do so. To serve them,
we keep a version that does not export the primary, but asserts
that the primary matches the first entry in its list.
Signed-off-by: Greg Farnum <greg@inktank.com>
This function does not (and never has!) used the raw vector, so remove it
and don't use a name which implies it is doing any sort of conversion.
Signed-off-by: Greg Farnum <greg@inktank.com>
These were the same except for a call to _raw_to_up_osds(). Move the
existing pg_to_up_acting_osds into a private function taking a pointer,
only fill in the up vector if it's a non-NULL pointer, and call it via
the obvious header implementations.
Signed-off-by: Greg Farnum <greg@inktank.com>
This was used only by SyntheticClient, and that wants get_pg_acting_primary()
anyway. Delete the easily-misused get_pg_primary() and switch.
Signed-off-by: Greg Farnum <greg@inktank.com>
Some of these look like what you should use for mapping and they absolutely
are not suitable for that. Make it clearer.
Signed-off-by: Greg Farnum <greg@inktank.com>
We've been using the role returned from this to determine if we're
the primary or not. Don't.
This is mostly about removing a few asserts; while in there I also
redirected some calls to use static dereference instead of going through
the osdmap lookup path.
Signed-off-by: Greg Farnum <greg@inktank.com>
When run in a shared environment ( as opposed as a machine created for
the purpose of running this test only ), it is important to cleanup
leftovers to avoid poluting the /tmp space. Create a common temporary
directory for all tmp files.
Signed-off-by: Loic Dachary <loic@dachary.org>
