Makefile.am: clean gcno and gcda files in "make clean"

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>

coverage: add helper script to get coverage for a local test

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>

mon: add all_exit and exit commands

all_exit makes each daemon exit(0), for gcov data collection.
exit causes cmon to do this.

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>

mds: allow mds to 'exit immediately'

This is temporary until shutting down cleans up well.

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>

mon: ceph tell mds * is a valid command

Previously this fell through and returned -EINVAL to the user.

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>

osd: add command to exit cleanly

This is required for gcov to work on daemons since the coverage data
is written atexit, and the function that writes the data is not
exported.

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>

configure: add option for building with gcov coverage support

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>

rados_sync: prefix user extended attributes

Start user extended attributes with USER_XATTR_PREFIX.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

librados: implement aio_flush

Implement a per-ioctx flush that blocks until all previously submitted
aio operations on the ioctx are safe.  Each aio gets a sequence number and
is put on a linked list attached to the ioctx.  The flush operation waits
for it to drain to the watermark set when flush is first called.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

crushtool: error out if uniform weights vary

Fixes: #1075
Signed-off-by: Sage Weil <sage@newdream.net>

osd: fix ScrubFinalizeWQ::_clear condition

Signed-off-by: Josh Durgin <josh.durgin@dreamhost.com>

debian: depend on libboost-dev >= 1.34

for statechart.  Partially fixes #1124.

Signed-off-by: Sage Weil <sage@newdream.net>

osd: don't leak Connection reference

Signed-off-by: Sage Weil <sage@newdream.net>

osd: ignore old/stale heartbeat messages

If we get heartbeat messages from old epochs from peers that are not
current, drop them and mark the connection down.  Even if they are peers
we _should_ have (because we haven't gotten a notify yet to learn about
a pg we should have but don't yet) we have a newer map epoch and will learn
about them shortly, reopening the connection.

Fixes: #1107
Signed-off-by: Sage Weil <sage@newdream.net>

osd: fix map sharing due to heartbeats

- share the map with the cluster addr
- use the new {note,get}_peer_epoch helpers to do it sanely
- don't share if we're booting; see 818fa33a661

Signed-off-by: Sage Weil <sage@newdream.net>

crushtool: add -v verbose for --test mode

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

hadoop: track Hadoop API changes

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

SimpleMessenger: allow multiple calls to shutdown

Fixes a case where radostool crashed on an error shutdown.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

common/Thread.h: const cleanup

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

Merge branch 'wip-obsync'

mkcephfs: pass config to osdmaptool

This lets OSDMap::create_simple() see g_conf.osd_pool_default_size when
creating the initial data, metadata, and rbd pools.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

drop useless cm.txt

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

osdmap: take default pool size from config

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

crushtool: update help

Signed-off-by: Sage Weil <sage@newdream.net>

obysnc: rgw target: validate all users

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

Merge branch 'wip-obsync'

mon: remove pg_temp mappings when we delete pools

Signed-off-by: Sage Weil <sage@newdream.net>

test-obsync: test sync directly from s3->rgw

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

crushtool: fix --add-item weight being zero when parent bucket(s) created

Signed-off-by: Sage Weil <sage@newdream.net>

obsync: fix bucket creation through rgw target

The rgw: target can now create buckets. Add a test.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

Merge branch 'stable'

test-obsync: test big objects, user-defined xattr

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

mkcephfs: set rdir for local mon setup

Fixes: #1113
Reported-by: Bernard Grymonpon <bernard@openminds.be>
Signed-off-by: Sage Weil <sage@newdream.net>

init-ceph: ssh

Another bell/whistle.

Signed-off-by: Sage Weil <sage@newdream.net>

obysnc: fix content-type on RGWStore

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

test-obsync: compare_directory now compares xattrs

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

ceph-pybind-test: test embedded NULLs in data

Test embedded nulls in rados data. Fix a bug in rados.Object.__str__

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

obsync: more fixes for RgwStore

* Fix content-type handling

* add vvprint and use it in Object::equals.

* support RgwStore::prefix

* more tests

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

pybind/rados: correctly return data with NULLs

Correctly handle returning data with embedded NULLs in it.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

pybind/rados.py: throw NoData on ENODATA

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

mds: fix canceled lock attempt

If client tries to lock a file, has to wait, and then cancels the attempt,
the client will send an unlock request to unwind its state.

 - the unlock now removes the waiting lock attempt from the wait list
 - when the lock request retries and finds it is no longer on the wait
   list it will fail.

Signed-off-by: Sage Weil <sage@newdream.net>

pybind/rados.py: rados.Object.key should be string

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

obysnc: RgwStore: make sure destination users exist

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

obsync: fix DST_OWNER

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

rgw: return EACCES if acl xattr doesn't exist

obsync: Add boto_retries, remove rgw_store.prefix

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

rados python bindings: handle xattrs with NULL

Handle extended attributes that contain NULL bytes correctly, rather
than treating everything as zero-terminated C strings.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

PG: fix race in _activate_committed

Previously, _activate_committed would access the osdmap epoch racing
with handle_osd_map's osdmap update.  This would allow a message to be
sent from a replica to the primary tagged with the same epoch as
last_warm_restart, though the event actually occured before
last_warm_restart.  Thus the primary would fail to ignore the event and
transition to crashed.

Signed-off-by: Samuel Just <samuel.just@dreamhost.com>

RgwStore: fix some ACL issues

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

test-obysnc.py: support librgw testing

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

Rename RadosStore to RgwStore

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

test-obsync: refactor a little bit

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

Proper ACL support for rados targets

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

mds: do not shift to EXCL or MIX while rdlocked

There was an old change in file_eval() that was allowing us to switch from
SYNC to MIX or EXCL while there were rdlocks, which either caused lots of
lock thrashing or could (I think) hang things up completely.  This was
from ea10a672, an ancient fix for something related that appears to have
taken out the rdlocked check by accident.

In my tests (one writer, one stat-er), this took things from long stalls
(up to 20 seconds) to very responsive stats.  Yay!

Fixes: #791
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

crushtool: clean up add-item a bit; don't add item to same bucket twice

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

crushtool: fix remove-item

Scan all buckets instead of doing a tree traverse.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

radosgw_admin: update clitest

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

mkcephfs.in: print out usage if no actions given

If the user didn't specify any actions, print out a usage message rather
than silently exiting.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

rgw: Fix RGWAccess::init_storage_provider

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

mkcephfs: error out on bad usage

Signed-off-by: Sage Weil <sage@newdream.net>

make: fix build for rgw

rgw_admin: clean warning

Merge commit 'origin/master' into rgw-multiuser

rgw_admin: add key create

rgw_admin: subuser and key removal

journaler: tolerate ENOENT when prezeroing

ENOENT is okay and expected.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

test_common.sh: skip rm before put

The rm before the put is unecessary and actually incorrect now.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

radostool: rados put should use write_full

If "rados put" uses write instead of write_full, the resulting object on
the server may be a mismash of old and new objects, if the old object
was longer than the new one. This is fairly counterintuitive behavior
for radostool, so remove it.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

Merge branch 'wip_ceph_context'

Create a libcommon service thread

Create a libcommon service thread. Use it to handle SIGHUP.

Handle it by means of a flag that gets set. Using a queue would raise
the complicated question of what to do when the queue was full.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

librados: len should be size_t

Unsigned, and size_t because it's a buffer size.

Fixes signedness warning in testrados.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

osd: add ability to explicitly mark unfound as lost

Instead of automatically marking unfound objects lost (once we've tried
every location we can think of), do it when the administator explicitly
says to.  This avoids marking things wrong incorrectly when there are
peering issues, and also allows the administrator to decide whether there
may be offline osds that are worth bringing online.

Signed-off-by: Sage Weil <sage@newdream.net>

osd: make automatically marking of unfound as lost optional

We may not want to do this automatically until we have more confidense in
the recovery code.  Even then, possible not.  In particular, the OSDs may
believe they have contact all possible homes for the data even though there
is some long-lost OSD that has the data on disk that if offline.

For now, we make the marking process explicit so that the administrator can
make the call.

Signed-off-by: Sage Weil <sage@newdream.net>

mds: clean up get_or_create_stray

Signed-off-by: Sage Weil <sage@newdream.net>

mds: initialize stray_index on startup

Signed-off-by: Sage Weil <sage@newdream.net>

Merge branch 'stable'

v0.28.1

librads, libceph: store CephContext

Don't use the global g_ceph_context. Instead, store the CephContext in
the structures provided by the library user.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

Add CephContext

A CephContext represents the context held by a single library user.
There can be multiple CephContexts in the same process.

For daemons and utility programs, there will be only one CephContext.
The CephContext contains the configuration, the dout object, and
anything else that you might want to pass to libcommon with every
function call.

Move some non-config things out of md_config_t and into CephContext.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

Split common_init_daemonize from common_init_finish

Split off common_init_daemonize from common_init_finish. cfuse is a
daemon that calls common_init_finish, but handles daemonization itself.
This fixes cfuse.

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

rgw_admin: make interface a bit more explicit

rgw: subuser permissions

mon: verify that crush max does not exceed osd max

- when injecting a new crushmap
- when adjusting osdmap max_osd

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

crushtool: add --reweight-item <name> <weight>

Reweight and individual item via crushtool.

Signed-off-by: Sage Weil <sage@newdream.net>

osdmaptool: fail --import-crush if crush max_devices > osdmap max_osd

Crush will spew non-deterministic badness if it walks off the end of
the osd_weight vector.

Signed-off-by: Sage Weil <sage@newdream.net>

common_init: don't init crypto until after fork

Get rid of the initialize-then-shutdown-crypto hack. We just initialize
crypto once, after it is safe to do so. There is now a single callback,
common_init_finish, which does the final stage of initialization,
including starting crypto and daemonization (if required.)

common_init_finish needs to be done before messenger::start().

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

ceph_crypto: add assert_init

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>

config: delete after new

Signed-off-by: Sage Weil <sage@newdream.net>

crush: fix signedness warnings

Signed-off-by: Sage Weil <sage@newdream.net>

rgw_admin: able to create multiple keys/subusers

crushtool: --remove-item name

Signed-off-by: Sage Weil <sage@newdream.net>

crush: fix tree bucket encoding

I wonder how long this has been broken!

Signed-off-by: Sage Weil <sage@newdream.net>

crush: fix tree weight accessor, decompile

Signed-off-by: Sage Weil <sage@newdream.net>

crushtool: default to hash 0 (rjenkins1)

Otherwise we get 255 which is undefined and get bad results!

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

rgw: user info structure supports multiple subusers and keys

osd: update last_epoch_clean in PG::Info::History::merge()

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

osd: small cleanup

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

osd: merge history when primary sends replica new pg info

This, among other things, lets us update last_epoch_started and
last_epoch_clean.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

osd: more heartbeat rework

A few things:
 - track Connection* instead of entity_inst_t for hb peers
 - we can only send maps over the cluster_messenger
   - if peer is still alive, do that
   - if peer is not, send dying MOSDPing ping with YOU_DIED flag

msgr: don't close close_on_empty until outgoing messages are acked

Otherwise, if we close the socket, we may lose in-flight data.

Signed-off-by: Sage Weil <sage.weil@dreamhost.com>