osd: abort on startup if we fail to bind to a port

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

ceph: fix "run_uml.sh" script

Last-minute cleverness prior to checkin broke the "run-uml.sh" script.
Rearange where a few definitions are done to make it work again.

Signed-off-by: Alex Elder <elder@dreamhost.com>

mds: fix uninitialized value in MClientLease::h

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

msgr: move operator<< for sockaddr_storage to msg_types.cc

tcp.{cc,h} aren't built/linked cleanly.

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

msgr: don't assert on socket(2) failure

This can happen if we're connecting to an invalid address.  Generate an
error message instead of crashing.

See #1942
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

msgr: uninline operator<< on sockaddr_storage

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

Merge branch 'stable'

ReplicatedPG: munge truncate_seq 1/truncate_size -1 to seq 0/size 0

Truncate with seq 1 and size -1 is a noop.

Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
Reviewed-by: Sage Weil <sage@newdream.net>

v0.40

ReplicatedPG: Update stat accounting for truncate during write

Signed-off-by: Samuel Just <samuel.just@dreamhost.com>
Reviewed-by: Josh Durgin <josh.durgin@dreamhost.com>

rgw: wrap cls_cxx_map_* with try/catch around decoding

rgw: bucket index creation and init in a single operation

librados: add ObjectOperation::exec

secret: move null check before strlen(key_name) deref

Coverity cid: 98
Signed-off-by: Sage Weil <sage@newdream.net>

osd: stat op, don't compare in memory state to object

might be that object is being created by the current compound request.

rgw: don't crash when copying a zero sized object

ReplicatedPG: Do a write even for 0 length operation

Otherwise, a 0 length write to an offset past the end of the file will
cause the internal accounting to reflect the full size of the file, but
not the file on disk.

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

ReplicatedPG: fix stat accounting error in CEPH_OSD_OP_WRITEFULL

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

qa/client/gen-1774.sh

Capture Alexandre's script for reproducing #1774 here for posterity, until
we write a properly harnessed test for this.  Currently, workunits can't
mount/unmount, and we don't have a way to make ceph-fuse drop it's cache.

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

client: avoid taking inode ref in case of nonexistent dir

Signed-off-by: Andrey Stepachev <octo@yandex-team.ru>
Signed-off-by: Sage Weil <sage@newdream.net>

Merge branch 'wip-makefile'

COPYING: note licenses for all files, not just the default

This (mostly) copies debian/copyright for now, but there are format
restrictions for that file.  Suggestions for a cleaner way to handle this
are welcome.  In the meantime, this is better...

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

debian/copyright: note acx_pthread.m4 license

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

Makefile: Add headers that were omitted in make dist and prevented tests from building

Signed-off-by: Kacper Kowalik (Xarthisius) <xarthisius@gentoo.org>

Makefile: Handle corner case of crypto++ correctly

i.e. use c++ while compiling, append to CRYPTO_LIBS instead of LIBS

Signed-off-by: Kacper Kowalik (Xarthisius) <xarthisius@gentoo.org>

Makefile: Use ACX_PTHREAD in configure.ac and resulting flags in src/Makefile.am

instead of hardcoded flags

Signed-off-by: Kacper Kowalik (Xarthisius) <xarthisius@gentoo.org>

Makefile: Add recent acx_pthread.m4 that has a fix for nostdlib issue.

See http://code.google.com/p/protobuf/issues/detail?id=188 for details

Signed-off-by: Kacper Kowalik (Xarthisius) <xarthisius@gentoo.org>

PG: gen_prefix should grab a map reference atomically

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

rgw-admin: add pool rm and pools list

rgw-admin: clean up unused commands

rados.py: avoid getting return value of void function

rados_ioctx_locator_set_key is void. The return value seems to have
been uninitialized, so the tests failed rarely.

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

client: start caching readdir results after readdir_start

Use upper_bound rather than lower_bound to compute the initial pd within
insert_trace, so that we don't attempt to remove it if it happens to be
in the same frag as the new reply.

Fixes: #1774
Signed-off-by: Alexandre Oliva <oliva@lsd.ic.unicamp.br>
Signed-off-by: Sage Weil <sage.weil@dreamhost.com>

monclient: fix resolve_addrs() call

This was broken in def36668a13459d9c0851e4d4da440a288f9a34f it looks like.
Passing uninitialized memory to resolve_addrs(), and needlessly
allocating a buffer.

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

resolve_addrs: return ipv4 and ipv6 addrs

Fixes: #1891
Signed-off-by: Sage Weil <sage@newdream.net>

ReplicatedPG: fix typo in stats accounting in _rollback_to

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

mon: allow specifying pg_num and pgp_num when creating new pools.

Right now this is only exposed via the monitor command interface:
osd pool create <poolname> [pg_num [pgp_num]]
but it can be expanded to other interfaces as appropriate.

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>

auth: Fix Doxygen warnings.

Match prototype and implementation argument names and types
(textually, that is use std:: prefix).

Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>

Fix several doxygen warnings, to minimize noise. Only changes comments.

Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>

librados: Make API docs use @note instead of @bug for now.

Asphyxiate doesn't yet support all of the Doxygen markup.

Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>

FileStore: assert on ENOSPC even for SETXATTR

Otherwise we can get corrupt object attributes on ext*.

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

mds: initiate monitor reconnect if beacon acks take too long

If it takes 2*mds_beacon_grace (default 30 seconds total) seconds
to get an ack back, maybe it's the monitor and not us. Try a reconnect,
which will just add the teensiest bit of load if we're wrong.

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>

mds: remove beacon_killer code.

This no longer does *anything* except print out
useless warning messages.

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>

osd: make less noise when filestore is already up to date

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

doc: add librados C aio example

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

doc: describe some rados_pool_stat_t members

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

doc: add librados pool creation defaults

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

doc: add short section on documenting code

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

doc: clarify librados return codes

Adding a second @returns for specific error codes makes the sphinx output more readable.

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

doc: @return -> @returns to match the sphinx output

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

doc: standardize rados_tmap_* docs

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

doc: fix rados_version todo formatting

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

doc: add a prefix to group names in librados.h

doxygen groups are in a global namespace.

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

doc: Put rados_ioctx_locator_set_key in a group so it can be cross-referenced

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

doc: move rados_ioctx_get_id to the pool group

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

doc: fix some typos in librados C API

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

doc: Switch doxygen integration from breathe to asphyxiate.

TODO: path of librados.h is now just the basename

TODO: no enum support for now

TODO: no @bug support for now

Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>

librados: Avoid using "crush_rule" as name of function argument.

"struct crush_rule" exists already, using the same identifier
confuses Doxygen.

Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>

doxygen: Use first sentence as brief description.

Signed-off-by: Tommi Virtanen <tommi.virtanen@dreamhost.com>

doc: add configuration and connecting to librados C api example

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

doc: add librados C api docs

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

ceph: add a new "run_uml.sh" script to manage running a UML client

This script is used to automate most of what's required to run a
User-Mode Linux (UML) instance.  This is mainly of interest for
ceph client developers who might benefit from the debugger access
that UML affords.  It was written for ceph development but isn't
really dependent on ceph.  It basically makes a few assumptions and
follows some conventions, and in doing so is able to encapsulate
most of the "tricky parts" of setting up to run a UML instance.

Signed-off-by: Alex Elder <elder@dreamhost.com>

rgw: adjust log level

rgw: some cleanup

rgw: only use plain PUT processor when !chunked_upload

Merge remote branch 'gh/wip-osd-retry-attempt'

Merge remote branch 'gh/wip-admin-socket'

perfcounters: fix unittest for new admin_socket interface

Broken by b389685afa1be00b5147855bf71c50042bfbfa6c.

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

Makefile: disable untitest_interval_tree

Segfaults. Valgrind errors. Accessing uninitialized memory.

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

unittest_interval_tree: make it compile

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

osd: read op should claim_append data instead of claim

rgw: remove object before writing both xattrs and data

otherwise we'll leak xattrs from previous incarnation

rgw: create plain processor for small objects

rgw: fix multipart PUT

latest revamp broke it, missed calling RGWPutObjProcessor::prepare(s)
where needed.

rgw: rearrange PutObj::execute()

groundwork for different handling of small object PUTs

rgw: different atomic handling for small objects

mon: fix uninitialized cluster_logger_registered

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

objecter: ignore replies from old request attempts

If we know the request attempt, ignore old attempts.

If we do not know the attempt (because the server is old), accept the
reply.  This could lead to doing some ACK callbacks we shouldn't in
extreme failure/recovery scenarios, but that is better than doing
the callbacks out of order.

Partially fixes: #1490
Signed-off-by: Sage Weil <sage@newdream.net>

osd: encode retry attempt in MOSDOp[Reply]

In addition to the boolean flag, also encode the exact retry attempt.

Return -1 if we don't know.

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

mon: document quorum_status, mon_status

Fixes: #1824
Signed-off-by: Sage Weil <sage@newdream.net>

mon: fix misplaced else

Broken by 435c29448a10ec343f5a2b7195d94c72de5b1a25.

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

Merge remote branch 'gh/wip-mon-timeouts'

ceph: speak new admin socket protocol

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

admin_socket: fix, extend admin_socket unit tests

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

admin_socket: string commands

Commands are strings.  Old __be32 works too.  'help' to list available
commands.

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

mon: elector needs to reset leader_acked on every election start

Otherwise you never reset the leader_acked after a failed
election attempt, so if mon 0 is available on the first round
but then fails, you never make progress!

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>

mon: instrument elector so you can stop participating in the quorum

Add new monitor commands "quorum exit" and "quorum enter" to use it.

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>

mon: kill client sessions when we're not in quorum

After a timeout of 2*mon_lease length (ie, two election rounds),
kill existing client sessions so they can reconnect to a
monitor that's (hopefully) remained in the quorum. Let any
new client sessions stick around for a mon_lease interval, then
do the same to them.

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>

OCF RA: fix variable name

debian: build ceph-resource-agents

admin_socket: refactor

Combine AdminSocketConfigObs with AdminSocket so that we can interact
with it via the cct.  Simpler class structure.  Less pointer indirection.

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

rbd: add a command to delete all snapshots of an image

This makes deleting images with many snapshots easier.

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

admin_socket: whitespace

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

librbd: don't remove an image that still has snapshots

Return -EBUSY instead. After the header is removed, the snapshots
can't be removed or read, so make sure they're gone before proceeding.

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

SimpleMessenger: clarify when ms_bind_ipv6 is used

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

qa: add a slightly more stressful anchortable test

This creates more than 8 links.

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>

qa: fix mdstable script for proper injectargs use.

This script is fairly primitive, but somebody might find it useful...

Signed-off-by: Greg Farnum <gregory.farnum@dreamhost.com>

rados: fix run-length option parsing for rados load-gen

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

Merge pull request #8 from kylemarsh/master

Remove cloudfiles requirement from obsync.

qa: load-gen-mix-small-long

30 minutes

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

obsync: make obsync run without cloudfiles installed

Cloudfiles probably shouldn't be a requirement for running obsync, so this
commit makes it optional.