kv_flat_btree_async.cc: remove unused variables

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

kv_flat_btree_async.cc: remove some unreachable break statements

Remove unreachable break statements after return calls in a switch.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

key_value_store/cls_kvs.cc: remove unused variable dupmap

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

bloom_filter.hpp: check for self assignment

Check for self assignment in operator= to avoid problems with
dynamic memory (cppcheck).

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

crush/builder.c: reduce scope of oldsize in crush_add_rule()

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

CrushCompiler.cc: remove duplicate if/else branches

Fix duplicate content of a if/else branch. Remove the
complete if/else around the content.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

common/secret.c: reduce the scope of some variables

Reduce the scope of error_buf to the place where it's used.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

pick_address.cc: pass const string by refrence

Pass 'const string networks' by reference to find_ip_in_subnet_list().

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

obj_bencher.cc: use static_cast instead of C-Style cast

Use static_cast<T>() instead of C-Style cast.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

config.cc: remove unused variable 'val' from parse_injectargs()

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

code_environment.cc: move variable ret in get_process_name()

Move 'ret' in get_process_name() to the correct block to suppress
warning about unused variable in cppcheck.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

ceph_context.cc: remove unused variable 'args' from do_command

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

cls_crypto.cc: remove some unused variables

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

cls_acl.cc: remove unused variables

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

common/armor.c: reduce scope of some variables

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

TextTable: pass const string by refrence to define_column()

Pass 'const string heading' by reference to define_column().

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

CephBroker.h: use static_cast instead of C-Style cast

Use static_cast<T>() instead of C-Style cast.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

SyntheticClient.cc: use static_cast instead of C-Style cast

Use static_cast<T>() instead of C-Style cast.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

Build: Change build to always use system leveldb

Dynamically link to the leveldb installed on the system rather than
statically linking ceph copy.  Remove the --with-system-leveldb config
option, and add a requirement for leveldb libraries for rpm and debian
packages.  Bug 3945.

Signed-off-by: Gary Lowell <gary.lowell@inktank.com>

ceph-test.install:  remove cepkfs-test.jar

This was merged prematurely to mast.

Signed-off-by: Gary Lowell <gary.lowell@inktank.com>

doc: fix rst for watch/notify internals

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

OpRequest: rmw_flags may actually be 0

If the operation includes only an invalid method,
the flags will be 0.

Reviewed-by: Yehuda Sadeh <yehuda@inktank.com>
Signed-off-by: Samuel Just <sam.just@inktank.com>

Merge branch 'next'

ceph-test.install:  Add libcephfs-test.jar.

Signed-off-by: Gary Lowell <gary.lowell@inktank.com>

Merge pull request #78 from Elbandi/master

Some fixes and functions for libcephfs and client

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

client: add some new functions: lchmod, fsetattr, lutime

Signed-off-by: Andras Elso <elso.andras@gmail.com>

client: allow change file owner or group only

Signed-off-by: Andras Elso <elso.andras@gmail.com>

client: use get_filehandle

Signed-off-by: Andras Elso <elso.andras@gmail.com>

client: fix log data

Signed-off-by: Andras Elso <elso.andras@gmail.com>

libcephfs: fix default parameters document for ceph_open_layout

Signed-off-by: Andras Elso <elso.andras@gmail.com>

Merge pull request #76 from dalgaaf/wip-da-sca-style

Fix some style related issues found by cppcheck

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

Merge remote-tracking branch 'gh/wip-4256'

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

PendingReleaseNotes: mention ceph-mon requiring the creation of data dir

Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>

mon: give useful errors to user when store is missing

We used to assert() instead, which didn't shed enough light on the cause
and could confuse the user into believing something *terrible* had
happened.

Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>

mon: Monitor: check for an argument to 'quorum' before reading the array

Fixes: #4218
Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>

mon: Monitor: adjust sync start criteria considering a peon's bootstrap

Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>

mon: Paxos: bootstrap if we have fallen behind when we start collecting

Fixes: #4256
Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>
Reviewed-by: Sage Weil <sage@inktank.com>

librbd: fix rollback size

The duplicate calls to get_image_size() and get_snap_size() replaced
by 5806226cf0743bb44eaf7bc815897c6846d43233 uncovered this. The first
call was using the currently set snap_id instead of the snapshot being
rolled back to.

Fixes: #4272
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
Reviewed-by: Sage Weil <sage@inktank.com>

msg: fix entity_addr_t::is_same_host() for IPv6

We weren't checking the memcmp return value properly!  Aie...

Backport: bobtail
Signed-off-by: Sage Weil <sage@inktank.com>

mon: fix other debug dumps

These should only kick in at high debug levels.  Moving any "work" between
the dout() and dendl lines puts them inside the magical hidden if block
that makes this code a no-op at low debug.

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

msgr: print dump before asserting (if that is enabled)

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

msgr: dump corrupt message to log (at high debug levels)

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

Merge remote-tracking branch 'gh/next'

osd: allow (some) log trim when degraded, but not during recovery

We allow some trim during degraded, although we keep more entries around to
improve our chances of a restarting OSD of doing log-based recovery.

Still disallow during recovery...

Signed-off-by: Sage Weil <sage@inktank.com>
Reviewed-by: Samuel Just <sam.just@inktank.com>

osd: restructure calc_trim

No functional change, except that we log more debug, yay!

Signed-off-by: Sage Weil <sage@inktank.com>
Reviewed-by: Samuel Just <sam.just@inktank.com>

osd: allow pg log trim during (non-classic) scrub

Chunky (and deep) scrub do not care about PG log trimming.  Classic scrub
still does.

Deep scrub can take a long time, so not trimming the log during that period
may eat lots of RAM; avoid that!

Might fix: #4179
Signed-off-by: Sage Weil <sage@inktank.com>
Reviewed-by: Samuel Just <sam.just@inktank.com>

Merge remote-tracking branch 'gh/wip_4257'

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

mon: avoid expensive dump unless log will gather

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

ceph_common.sh: tolerate missing mds, mon, osds in conf

With set -e this seems to fail (at least on some machines) if, say, there
is no MDS in the conf file.  This fixes it.

Tested-by: Mark Nelson <mark.nelson@inktank.com>
Signed-off-by: Sage Weil <sage@inktank.com>

Merge branch 'wip-4119'

Fix for bug 4119r, create directories under /var/lib/ceph.

testing: updating hadoop-internal test

Small tweaks to the hadoop-internal test
to better use existing environment varaibles
and in response to the recent teuthology
changes.

Signed-off-by: Joe Buck <jbbuck@gmail.com>
Reviewed-by: Noah Watkins <noahwatkins@gmail.com>

CephBroker.cc: fix condition

from cppcheck:
 Fix Suspicious condition (assignment + comparison); Clarify
 expression with parentheses.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

test: removing dead code

Signed-off-by: Joe Buck <jbbuck@gmail.com>

test: updating libcephfs-java tests

Recent changes in teuthology necessitate
changes to libcephfs-java test case.

Signed-off-by: Joe Buck <jbbuck@gmail.com>

Client.cc: reduce the scope of some variables

Reduce the scope of some variables.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

Client.cc: use static_cast instead of C-Style cast

Use static_cast<T>() instead of C-Style cast.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

Merge pull request #75 from alram/master

Typo: pool should be root in rados/operations/crush-map.rst

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

Paxos.h: pass string name function parameter by reference

Pass 'const string name' function parameter by reference.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

Typo: pool should be root in rados/operations/crush-map.rst

Signed-off-by: Alexandre Marangone <alexandre.marangone@inktank.com>

Merge pull request #73 from dalgaaf/wip-da-sca-cppcheck-emptiness

fix more emptiness checks

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

Merge pull request #74 from dalgaaf/wip-da-sca-misc

mds/journal.cc: fix assert in replay()

Let's see how many bugs this has been hiding... :)

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

osd/ReplicatedPG.cc: remove unused variables

Remove unused variable declarations.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

osd_types.cc: add parenthesis around expression

Add parenthesis around expressions like (x && y & z).

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

OSDMap.cc: add parenthesis around expression

Add parenthesis around expressions like (x && y & z).

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

PGMonitor.cc: add parenthesis around expression

Add parenthesis around expressions like (x && y & z).

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

Client.cc: add parenthesis around expression

Add parenthesis around expressions like (x && y & z).

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

CephxProtocol.cc: catch end_of_buffer by reference

Follow 'Throw by value, catch by reference' to void copying and
potential slicing the exception.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

CephxProtocol.cc: catch error by reference

Follow 'Throw by value, catch by reference' to void copying and
potential slicing the exception.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

crushtool.cc: remove unneeded function declaration

Remove unneeded function declarations (data_analysis() is unused and
usage() is not needed).

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

mds/journal.cc: fix assert in replay()

Fix assert in replay(), check if mds->sessionmap.version is cmapv
instead of assign cmapv to  mds->sessionmap.version. Add missing
second '='.

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

KeyValueDBMemory.cc: use empty() instead of size() == 0

Use empty() since it should be prefered as it has, following the
standard, a constant time complexity regardless of the containter
type. The same is not guaranteed for size().

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

ReplicatedPG.cc: use empty() instead of size() == 0

Use empty() since it should be prefered as it has, following the
standard, a constant time complexity regardless of the containter
type. The same is not guaranteed for size().

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

FileStore.cc: use if(!empty()) instead of if(size())

Use empty() since it should be prefered as it has, following the
standard, a constant time complexity regardless of the containter
type. The same is not guaranteed for size().

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

mon_store_converter.cc: use empty() instead of size()

Use empty() since it should be prefered as it has, following the
standard, a constant time complexity regardless of the containter
type. The same is not guaranteed for size().

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

Paxos.cc: use empty() instead of size()

Use empty() since it should be prefered as it has, following the
standard, a constant time complexity regardless of the containter
type. The same is not guaranteed for size().

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

Monitor.h: use empty() instead of !size()

Use empty() since it should be prefered as it has, following the
standard, a constant time complexity regardless of the containter
type. The same is not guaranteed for size().

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

Monitor.cc: use empty() instead of size()

Use empty() since it should be prefered as it has, following the
standard, a constant time complexity regardless of the containter
type. The same is not guaranteed for size().

Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>

ceph.spec.in:  Create placeholder directorys under /var/lib/ceph (Bug 4119)

Signed-off-by: Gary Lowell <gary.lowell@inktank.com>

configure.ac: Add test for c++ compiler.

The AC_PROG_CXX macro sets a flag if a C++ compiler is found
but does not fail if one is not found, it left to application
to test the flags as needed. This fix will issue an error
when a c++ compiler is not found.  Bug 3955.

Signed-off-by: Gary Lowell <gary.lowell@inktank.com>
it's not installed, this fix adds an error message for a

test_lock_fence.sh, rbdrw.py: rbd lock/fence test

qa/workunits/rbd/test_lock_fence.sh runs using test/rbdrw.py

rbdrw.py creates an image, locks it, and runs an I/O loop;
test_lock_fence.sh runs it, waits, and then blacklists that client,
which causes rbdrw.py to get ESHUTDOWN on operations thereafter.
Currently doesn't work with rbd caching enabled.

rbd.py gets new exception type for ESHUTDOWN

Fixes: #3190
Signed-off-by: Dan Mick <dan.mick@inktank.com>
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>

Merge remote-tracking branch 'gh/wip-4249-master'

Merge remote-tracking branch 'gh/wip-4249' into next

Merge remote-tracking branch 'gh/wip-4252'

mon: PaxosService: remove lingering uses of paxos getters and wait methods

We should use the PaxosServices getters, setters, and wait methods when and
wherever possible.  These must have fallen through the cracks during the
merge.

Signed-off-by: Joao Eduardo Luis <joao.luis@inktank.com>

systest: restrict list error acceptance

Only ignore errors after the midway point if the midway_sem_post is
defined.

Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
(cherry picked from commit 5b24a68b6e7d57bac688021b822fb2f73494c3e9)

systest: fix race with pool deletion

The second test have pool deletion and object listing wait on the same
semaphore to connect and start. This led to errors sometimes when the
pool was deleted before it could be opened by the listing process. Add
another semaphore so the pool deletion happens only after the listing
has begun.

Fixes: #4147
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>
(cherry picked from commit b0271e390564119e998e18189282252d54f75eb6)

Merge remote-tracking branch 'gh/wip-4147'

doc: Added subnet example and verbiage to network settings.

fixes: #4049

Signed-off-by: John Wilkins <john.wilkins@inktank.com>

doc: Added content to remove REJECT rules from iptables.

Signed-off-by: John Wilkins <john.wilkins@inktank.com>

test_rbd: move flatten tests back into TestClone

They need the same setup, and it's easy enough to run specific
subtests. Making them a separate subclass accidentally duplicated
tests from TestClone.

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

qa: enable watch-notify dependent test

This works now that watch-notify has been reworked a bit.

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

test_rbd: close image before removing it

This error was masked before by watch notify not differentiating
between watches from the same client with different cookies.
Reopen the image at the end of this test so teardown works.

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

doc: Added a small ref section for osd config reference.

Signed-off-by: John Wilkins <john.wilkins@inktank.com>

doc: Updated osd configuration reference.

Signed-off-by: John Wilkins <john.wilkins@inktank.com>

systest: restrict list error acceptance

Only ignore errors after the midway point if the midway_sem_post is
defined.

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

systest: fix race with pool deletion

The second test have pool deletion and object listing wait on the same
semaphore to connect and start. This led to errors sometimes when the
pool was deleted before it could be opened by the listing process. Add
another semaphore so the pool deletion happens only after the listing
has begun.

Fixes: #4147
Signed-off-by: Josh Durgin <josh.durgin@inktank.com>

qa: output rados test names as they run

So we don't have to figure out which test is running from the output,
which can be difficult with the system tests.

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

PG,ReplicatedPG: use pg_has_reset_since to discard old async events

Local async events are obsolete if the pg is deleting or if the
epoch at which they were created is prior to last_peering_reset.

Signed-off-by: Samuel Just <sam.just@inktank.com>

PG::build_scrub_map: detect race with peering via last_peering_reset

Signed-off-by: Samuel Just <sam.just@inktank.com>

ReplicatedPG::C_OSD_CommittedPushedObject: use intrusive_ptr for pg

Signed-off-by: Samuel Just <sam.just@inktank.com>