test: run-rbd-tests test cmake fixup

Signed-off-by: Mykola Golub <mgolub@mirantis.com>

Merge pull request #9731 from badone/wip-doc-ceph-authtool-print-fix

doc: ceph-authtool man page option is --print-key not --print

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #10124 from trociny/wip-cmake-tests

test: cmake related fixups for rbd tests

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
Reviewed-by: Kefu Chai <kchai@redhat.com>

test: fix run-rbd-tests test for cmake

Signed-off-by: Mykola Golub <mgolub@mirantis.com>

test: fix run-rbd-valgrind-unit-tests when built with cmake

Signed-off-by: Mykola Golub <mgolub@mirantis.com>

test: ceph_test_librbd does not need rados_test_stub

Signed-off-by: Mykola Golub <mgolub@mirantis.com>

Merge pull request #10117 from xiexingguo/xxg-wip-bluestore-2016-07-04

os/bluestore: fix incorrect pool decoding of bnode

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #9755 from dx9/wip-ns-name-uncompress

common/dns_resolve: use ns_name_uncompress instead of ns_name_ntop

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #9766 from dx9/wip-autotools-arch

configure.ac: Use uname instead of arch.

Reviewed-by: Willem Jan Withagen <wjw@digiware.nl>

Merge pull request #9772 from tchaikov/wip-close-on-exec

msgr: set close on exec flag

Reviewed-by: Haomai Wang <haomai@xsky.com>

Merge pull request #9820 from dx9/wip-fix-res-query-detect

configure.ac: fix res_query detection

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #9845 from dx9/wip-config-basename

common: remove basename() dependency

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #9969 from liewegas/wip-assert

include/assert: clean up ceph assertion macros

Reviewed-by: Greg Farnum <gfarnum@redhat.com>

Merge pull request #9983 from SUSE/wip-cmake-fix-fcgi-include-dir

cmake: Fix FCGI include directory

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #10093 from tchaikov/wip-16525

crush: reset bucket->h.items[i] when removing tree item

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #9330 from dachary/wip-15176-partprobe

ceph-disk: partprobe should block udev induced BLKRRPART

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #9394 from xiexingguo/xxg-wip-batchpgstates

mon/PGMonitor: batch filter pg states; add sanity check

Reviewed-by: Kefu Chai <kchai@redhat.com>

os/bluestore: fix incorrect pool decoding of bnode

"pool" is a pointer, and we shall alter its value instead.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>

ceph-disk: partprobe should block udev induced BLKRRPART

Wrap partprobe with flock to stop udev from issuing BLKRRPART because
this is racy and frequently fails with a message like:

    Error: Error informing the kernel about modifications to partition
    /dev/vdc1 -- Device or resource busy.  This means Linux won't know about
    any changes you made to /dev/vdc1 until you reboot -- so you shouldn't
    mount it or use it in any way before rebooting.

Opening a device (/dev/vdc for instance) in write mode indirectly
triggers a BLKRRPART ioctl from udev (starting version 214 and up)
when the device is closed (see below for the udev release note).

However, if udev fails to acquire an exclusive lock (with
flock(fd, LOCK_EX|LOCK_NB); ) the BLKRRPART ioctl is not issued.

https://github.com/systemd/systemd/blob/045e00cf16c47bc516c0823d059b7548f3ce9c7c/src/udev/udevd.c#L1042

Acquiring an exclusive lock before running the process that opens the
device in write mode is therefore an effective way to control this
behavior.

git clone git://anonscm.debian.org/pkg-systemd/systemd.git
systemd/NEWS:
CHANGES WITH 214:

  * As an experimental feature, udev now tries to lock the
       disk device node (flock(LOCK_SH|LOCK_NB)) while it
       executes events for the disk or any of its partitions.
       Applications like partitioning programs can lock the
       disk device node (flock(LOCK_EX)) and claim temporary
       device ownership that way; udev will entirely skip all event
       handling for this disk and its partitions. If the disk
       was opened for writing, the close will trigger a partition
       table rescan in udev's "watch" facility, and if needed
       synthesize "change" events for the disk and all its partitions.
       This is now unconditionally enabled, and if it turns out to
       cause major problems, we might turn it on only for specific
       devices, or might need to disable it entirely. Device Mapper
       devices are excluded from this logic.

Fixes: http://tracker.ceph.com/issues/15176
Signed-off-by: Marius Vollmer <marius.vollmer@redhat com>
Signed-off-by: Loic Dachary <loic@dachary.org>

Merge pull request #10077 from yehudasa/wip-16494

rgw: fix for issue 16494

Merge pull request #10064 from BlaXpirit/brag-py3

Port ceph-brag to Python 3 (+ small fixes)

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #10115 from runsisi/ss-wip-fix-rpm-spec

ceph.spec.in: fix rpm package building error

Reviewed-by: Kefu Chai <kchai@redhat.com>

ceph.spec.in: fix rpm package building error as follows:

      error: Installed (but unpackaged) file(s) found:
      /usr/lib/udev/rules.d/60-ceph-by-parttypeuuid.rules

Signed-off-by: runsisi <runsisi@zte.com.cn>

Merge pull request #9885 from dachary/wip-16351-by-parttypeuuid

udev: always populate /dev/disk/by-parttypeuuid

Reviewed-by: Alexandre Derumier <aderumier@odiso.com>

Merge pull request #10095 from liewegas/wip-bluestore-trim

os/bluestore: trim cache on reads

Reviewed-by: xie xingguo <xie.xingguo@zte.com.cn>

Merge pull request #10080 from xiexingguo/xxg-wip-add-compress-ratio

os/bluestore: improve required compression threshold

Reviewed-by: Sage Weil <sage@

msg/simple: set close on exec on server sockets

mds execv() when handling the "respawn" command, to avoid fd leakage,
and enormous CLOSE_WAIT connections after respawning, we need to set
FD_CLOEXEC flag for the socket fds.

Fixes: http://tracker.ceph.com/issues/16390
Signed-off-by: Kefu Chai <kchai@redhat.com>

msg/async: set close on exec on server sockets

mds execv() when handling the "respawn" command, to avoid fd leakage,
and enormous CLOSE_WAIT connections after respawning, we need to set
FD_CLOEXEC flag for the socket fds.

Fixes: http://tracker.ceph.com/issues/16390
Signed-off-by: Kefu Chai <kchai@redhat.com>

os/bluestore: add compression required ratio to enable/disable compression

Require the net gain of compression at least to be at a specified ratio,
otherwise we don't compress.

Ask for compressing at least 12.5% off, by default.

This is for the sake of performance because if the compression turns out
to be meaningless(saving little space), we can simply shut it down, as we
know the compression/decompression can be rather CPU-consuming.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>

Merge pull request #10096 from ceph/wip-dang-cmake

debian: set libexec dir to correct value as autotools did

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #9941 from weiqiaomiao/wqm-wip-fetchobj

rgw: fix memory leak in RGWRados::fetch_remote_obj()

lgtm

debian - set libexec dir to correct value as autotools did

Signed-off-by: Daniel Gryniewicz <dang@redhat.com>

os/bluestore: trim cache on reads too

We were only trimming in the write path, which would make memory explode
on a read-only workload.

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

Merge pull request #10082 from chhabaramesh/bitalloc_warn

os/bluestore: Remove bit alloc Woverloaded-virtual warnings

Merge pull request #10068 from jcsp/wip-cephfs-doc-update

doc: update CephFS "early adopters" info

Reviewed-by: Sage Weil <sage@redhat.com>
Reviewed-by: Greg Farnum <gfarnum@redhat.com>

Merge pull request #9584 from chiehtu/fix-16188

rgw: fix Swift API returns double space usage

Reviewed by: Pritha Srivastava <prsrivas@redhat.com>

Merge pull request #10000 from weiqiaomiao/wqm-wip-add-userlist-cli

rgw: add "radosgw-admin user list" cli

verified

Merge pull request #9838 from xiexingguo/xxg-wip-fix-synthetic-client

client: misc fixes

Reviewed-by: John Spray <john.spray@redhat.com>

Merge pull request #10092 from ceph/wip-dang-cmake

cmake: fixes

Reviewed-by: Kefu Chai <kchai@redhat.com>

Merge pull request #10028 from ceph/wip-cmake-readme-fixes

cmake: Removed README.cmake.md, edited README.md

crush: reset bucket->h.items[i] when removing tree item

* crush: so we don't see the reference after the removing, this keeps
  check_item_loc() happy, and move_bucket() use check_item_loc() to see if
  the removed bucket disappears after the removal.
* test: also add unittest_crush_wrapper::CrushWrapper.insert_item

Fixes: http://tracker.ceph.com/issues/16525
Signed-off-by: Kefu Chai <kchai@redhat.com>

init-ceph - use sbindir, rather than hard coding sbin

Signed-off-by: Daniel Gryniewicz <dang@redhat.com>

CMake - add libexecdir to the replaced variables

Signed-off-by: Daniel Gryniewicz <dang@redhat.com>

Merge pull request #10081 from tchaikov/wip-16556

cmake: remove unnecessary linked libs from libcephfs

Reviewed-by: John Spray <john.spray@redhat.com>

Merge pull request #10083 from trociny/wip-16539

rbd-mirror: remove ceph_test_rbd_mirror_image_replay test case

Reviewed-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #10065 from dillaman/wip-16558

rbd-mirror: gracefully fail if object map is unavailable

Reviewed-by: Mykola Golub <mgolub@mirantis.com>

cmake: remove unnecessary linked libs from libcephfs

* some of the libs shares the same .cc which has static C++ variables. if
  we link against the different libs sharing the same static C++
  variables, and the dtor of the C++ variables has side-effects, among
  other things, deallocates a memory chunk. then, we are in the trouble of
  double free. so "osd" lib is removed.
* some of the libs are referenced by the linked lib, so no need to link
  against them again. for example, BLKID_LIBRARIES are linked by
  libcommon, so we can remove it from the linked libs list.
* and lib "os" and "cls_references_objs" are not used by libcephfs at all,
  so remove them.

Fixes: http://tracker.ceph.com/issues/16556
Signed-off-by: Kefu Chai <kchai@redhat.com>

Remove bit alloc Woverloaded-virtual warnings

Signed-off-by: Ramesh Chander <Ramesh.Chander@sandisk.com>

rgw: fix double counting in RGWRados::update_containers_stats()

Fixes: http://tracker.ceph.com/issues/16188
Signed-off-by: Albert Tu <albert.t@inwinstack.com>

rgw: forward input data when forwarding set_bucket_version to master

Fixes: http://tracker.ceph.com/issues/16494
Needed to keep input data around to be forwarded correctly. Also, master
does not send any data back, so don't try to parse anything.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

rgw: set args when rebuilding req_info

Was missing args, so it didn't sign requests with subresources
correctly when forwarding.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>

doc/cephfs: rename early-adopters to best-practices

...and give it a link in the TOC (in addition
to where it is referenced from the top of index.rst)

Signed-off-by: John Spray <john.spray@redhat.com>

doc/cephfs: remove some scary warnings

...and restructure the "early adopters" page
into a "best practices" guide.  Early adopters
are now just adopters :-)

Signed-off-by: John Spray <john.spray@redhat.com>

cmake: Removed README.cmake.md, edited README.md

Deleted README.cmake.md and put any helpful
information it had into README.md.

Also a added note about which tests get executed
by ctest.

Signed-off-by: Ali Maredia <amaredia@redhat.com>

Merge pull request #9891 from diluga/rgw_socket_backlog_with_sign

RGW : setting socket backlog for via ceph.conf

Merge pull request #9852 from dx9/wip-rgw-ldap-link

rgw-ldap: add ldap lib to rgw lib deps based on build config

Merge pull request #9931 from Yan-waller/yj-wip-rgwmainccfix

rgw: remove uneccessary temporary variable and fix some indents

cosmetic changes only--build verified, f23

Merge pull request #9979 from Yan-waller/yj-wip-rgwrests3cc

rgw: fix potential memory leaks and remove unnecessary assignment

lgtm--should backport

Merge pull request #10031 from yehudasa/wip-16530

rgw: finish error_repo cr in stop_spawned_services()

Reviewed-by: Casey Bodley <cbodley@redhat.com>

rbd-mirror: remove ceph_test_rbd_mirror_image_replay test case

Fixes: http://tracker.ceph.com/issues/16539
Signed-off-by: Mykola Golub <mgolub@mirantis.com>

Merge pull request #9985 from Yan-waller/yj-wiph-rgwopccfixstring

rgw: use string::npos while 'string::find()' find nothing

lgtm--build verified, f23

Merge pull request #9898 from liewegas/wip-bluefs-async-flush

os/bluestore/BlueFS: do not hold internal lock while waiting for IO

Merge pull request #10002 from Yan-waller/yj-wip-rgwradosccremoveunused

rgw: remove uneccessary temporary variable 'r'

lgtm--build verified, f23

Merge pull request #10067 from liewegas/wip-do-cmake

do_cmake.sh: set up initial plugin dir

os/bluestore/BlueFS: debug statements to check log_fnode size in super block

Signed-off-by: Varada Kari <varada.kari@sandisk.com>

os/bluestore/Bluefs: Unit test cases for bluefs async flush

Signed-off-by: Varada Kari <varada.kari@sandisk.com>

os/bluestore/BlueFS: Make fysnc return the status

Signed-off-by: Varada Kari <varada.kari@sandisk.com>

os/bluestore/BlueFS: make _sync_and_flush_log smarter

If we know what event we need to wait for, only wait long enough for it
to flush.  This helps the situation where another thread flushed what we
needed, and more dirty stuff was added to log_t, but we don't need to
wait for that too for our caller to be happy.

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

os/bluestore: drop lock while we flush the log

Handle cases where we have multiple racing threads trying to flush the
log by only allowing one concurrent log flush to be in progress at a time,
and behave if, after flushing, there are no more dirty records to flush.

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

os/bluestore/BlueFS: use existing IOContext for _write_super

We need to be careful with IOContext destruction racing with the bdev
io completion thread.

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

os/bluestore/BlueFS: pass unique_lock down through _fsync and log flush

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

os/bluestore/BlueFS: carry lock for mkfs

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

os/bluestore/KernelDevice: rebuild buffer if we exceed IOV_MAX

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

common/buffer: assert that we do not exceed IOV_MAX in prepare_iov

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

unittest_bluefs: don't silence log

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

os/bluestore/BlueFS: _flush_bdev -> flush_bdev

It is safe to call without a lock.

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

os/bluestore/BlueFS: drop lock while waiting for user io to complete

_flush_wait is safe to call without a lock, as long as our reference is
stable.  Rename it wait_for_aio() to be more clear about what it does and
the fact that it doesn't require a lock.

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

os/bluestore/BlueFS: track dirty by log_seq, log_seq_stable

Note when we dirty a file, and clean it only if that seq has been
committed.  Currently this is always the case because we don't drop the
lock, but that will change shortly.

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

os/bluestore/BlueFS: do not dirty log file when size changes

We only need to dirty the log file if the allocate changes.  Replay is
smart enough to learn the file size as it goes.

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

os/bluestore/BlueFS: track log_seq_stable

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

os/bluestore/BlueFS: _flush_log -> _flush_and_sync_log

Be accurate.

No functional change.

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

Merge pull request #10018 from ifed01/wip-bluestore-squeeze-lextents

os/bluestore: apply "small encoding" for onode_t::extents map

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #10059 from tchaikov/wip-cmake

cmake changes

Reviewed-by: Ali Maredia <amaredia@redhat.com>

Merge pull request #10035 from Yan-waller/yj-wip-rgwradoscc-0630

rgw: fix the missing return value and remove temporary variables

lgtm, build verified f23

cmake: install rbd-replay-many and its manpage

rpm packaging needs it explicitly.

Signed-off-by: Kefu Chai <kchai@redhat.com>

Merge pull request #10029 from wjwithagen/patch-1

msg/async/net_handler.cc: make it more compatible with BSDs

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #10030 from stiopaa1/msg_asyncmess_removeCodeDup

msg/AsyncMessenger.cc: remove code duplication

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #10066 from Ved-vampir/isal_submodule

isa-l: add isa-l library as a submodule

Merge pull request #9999 from xiexingguo/xxg-wip-bluestore-2016-06-29

os/bluestore: misc fixes

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #9946 from xiexingguo/xxg-wip-bluestore-cache

os/bluestore: make trim() of 2Q cache more fine-grained

Reviewed-by: Sage Weil <sage@redhat.com>

Merge pull request #9968 from ceph/wip-no-start-update

rpm: Do not start targets on update

Reviewed-by: Nathan Cutler <ncutler@suse.com>

cmake: add compressor_plugins as dep for libos

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

common/PluginRegistry: try $plugin_dir/$foo if $plugin_dir/$type/$foo fails

This is helpful for running out of a build dir, where e.g. cmake puts
stuff in lib/* and type separated into subdirs by type.

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

do_cmake.sh: create initial ceph.conf setting dirs

This is enough for unit tests to run.

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

isa-l: add isa-l library as a submodule
Signed-off-by: Alyona Kiseleva <akiselyova@mirantis.com>

Merge pull request #9935 from xiexingguo/xxg-wip-make-zone-configurable

os/bluestore: fix memory leak during bit_alloc testing

Reviewed-by: Ramesh Chander <Ramesh.Chander@sandisk.com>

brag: Use Counter from standard library if available

Signed-off-by: Oleh Prypin <oleh@pryp.in>

rbd-mirror: gracefully fail if object map is unavailable

If the exclusive lock was lost due to a watch failure from an
overloaded cluster, gracefully abort the image sync.

Fixes: http://tracker.ceph.com/issues/16558
Signed-off-by: Jason Dillaman <dillaman@redhat.com>

brag: Make usage of bytes vs unicode compatible with Python 3

Signed-off-by: Oleh Prypin <oleh@pryp.in>

brag: Fix relative import for Python 3 compatibility

Signed-off-by: Oleh Prypin <oleh@pryp.in>