]> git.apps.os.sepia.ceph.com Git - ceph.git/log
ceph.git
10 years agorm old ceph-deploy reference 3924/head
Ken Dreyer [Tue, 10 Mar 2015 16:23:30 +0000 (10:23 -0600)]
rm old ceph-deploy reference

There's no need to refer to this old version of ceph-deploy. v1.1.3 is
circa June 2013, and since that time, we have shipped much newer version
of ceph-deploy (1.5.22 at the time of this writing). We should simply
state the behavior of modern ceph-deploys here.

Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
10 years agoMerge pull request #3887 from stiopaa1/feature10891
Kefu Chai [Tue, 10 Mar 2015 14:16:29 +0000 (22:16 +0800)]
Merge pull request #3887 from stiopaa1/feature10891

mon: add "osd pool get $pool all" command

Reviewed-by: Kefu Chai <kchai@redhat.com>
10 years agoMerge remote-tracking branch 'gh/hammer'
Sage Weil [Tue, 10 Mar 2015 13:24:59 +0000 (06:24 -0700)]
Merge remote-tracking branch 'gh/hammer'

Conflicts:
src/gmock

10 years agoMerge pull request #3901 from dachary/wip-10922-ceph-deploy-documentation
Danny Al-Gaaf [Tue, 10 Mar 2015 09:53:56 +0000 (10:53 +0100)]
Merge pull request #3901 from dachary/wip-10922-ceph-deploy-documentation

doc: preparing a device implies activate

Reviewed-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
10 years agoMerge pull request #3907 from dachary/wip-test-delays
Danny Al-Gaaf [Tue, 10 Mar 2015 09:51:05 +0000 (10:51 +0100)]
Merge pull request #3907 from dachary/wip-test-delays

tests: increase the delay waiting for a cluster event

Reviewed-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
10 years agoMerge pull request #3914 from ktdreyer/wip-doc-rgw-apache-log-perms
Danny Al-Gaaf [Tue, 10 Mar 2015 09:47:51 +0000 (10:47 +0100)]
Merge pull request #3914 from ktdreyer/wip-doc-rgw-apache-log-perms

doc: rm Apache log permissions instructions

Reviewed-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
10 years agoMerge pull request #3917 from dachary/wip-rhel6-doc
Kefu Chai [Tue, 10 Mar 2015 08:02:12 +0000 (16:02 +0800)]
Merge pull request #3917 from dachary/wip-rhel6-doc

doc: rhel6.5 is rhel6

Reviewed-by: Kefu Chai <kchai@redhat.com>
10 years agoMerge pull request #3890 from XunleiGFS/fix_C_TwoContexts
Kefu Chai [Tue, 10 Mar 2015 04:16:03 +0000 (12:16 +0800)]
Merge pull request #3890 from XunleiGFS/fix_C_TwoContexts

osdc: fix a memory leak in C_TwoContexts

Reviewed-by: Kefu Chai <kchai@redhat.com>
10 years agoMerge pull request #3845 from ceph/wip-wn-rgw-hammer
Sage Weil [Tue, 10 Mar 2015 02:49:21 +0000 (19:49 -0700)]
Merge pull request #3845 from ceph/wip-wn-rgw-hammer

Wip wn rgw hammer

Reviewed-by: Sage Weil <sage@redhat.com>
10 years agoMerge pull request #3913 from yuyuyu101/wip-async-fix-throttle-deadlock
Sage Weil [Tue, 10 Mar 2015 02:48:11 +0000 (19:48 -0700)]
Merge pull request #3913 from yuyuyu101/wip-async-fix-throttle-deadlock

Wip async fix throttle deadlock

Reviewed-by: Sage Weil <sage@redhat.com>
10 years agoMerge pull request #3911 from athanatos/wip-11057
Sage Weil [Tue, 10 Mar 2015 02:43:02 +0000 (19:43 -0700)]
Merge pull request #3911 from athanatos/wip-11057

ReplicatedPG: block write on degraded object if there are waiters

Reviewed-by: Sage Weil <sage@redhat.com>
10 years agodoc: rhel6.5 is rhel6 3917/head
Loic Dachary [Tue, 10 Mar 2015 01:06:10 +0000 (02:06 +0100)]
doc: rhel6.5 is rhel6

RHEL 6.5 is actually  http://ceph.com/rpm-giant/rhel6/

Signed-off-by: Loic Dachary <ldachary@redhat.com>
10 years agoMerge pull request #3909 from dachary/wip-11068-debian-jessie
Loic Dachary [Mon, 9 Mar 2015 22:40:04 +0000 (23:40 +0100)]
Merge pull request #3909 from dachary/wip-11068-debian-jessie

deb: add zlib1g-dev to Build-Depends for Debian/jessie

Reviewed-by: Yann Dupont <yann@objoo.org>
10 years agomon/OSDMonitor: refactor and new key "all" for osd pool get command 3887/head
Michal Jarzabek [Thu, 5 Mar 2015 18:01:05 +0000 (18:01 +0000)]
mon/OSDMonitor: refactor and new key "all" for osd pool get command

This will only output all the values applicable to a given type of pool.
So for example for a pool that is not a tier pool values like HIT_SET_TYPE,
HIT_SET_PERIOD, HIT_SET_COUNT etc. will be ignored.

Fixes: #10891
Signed-off-by: Michal Jarzabek <stiopa@gmail.com>
10 years agodoc: rm Apache log permissions instructions 3914/head
Ken Dreyer [Mon, 9 Mar 2015 19:12:25 +0000 (13:12 -0600)]
doc: rm Apache log permissions instructions

On all distros, Apache should be able to write to its own log directory.
If not, that is a severe distro bug that should be reported/fixed.

Remove the references to running "chown" on /var/log/httpd or
/var/log/apache2.

Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
10 years agodoc/release-notes: update notes for v0.80.9
Sage Weil [Mon, 9 Mar 2015 17:56:11 +0000 (10:56 -0700)]
doc/release-notes: update notes for v0.80.9

Signed-off-by: Sage Weil <sage@redhat.com>
10 years agoMerge pull request #3608 from guangyy/wip-rgw-quota-threads
Yehuda Sadeh [Mon, 9 Mar 2015 16:39:58 +0000 (09:39 -0700)]
Merge pull request #3608 from guangyy/wip-rgw-quota-threads

rgw: make quota/gc thread configurable for starting

Reviewed-by: Yehuda Sadeh <yehuda@redhat.com>
10 years agoMerge pull request #3795 from rzarzynski/wip-10938
Yehuda Sadeh [Mon, 9 Mar 2015 16:37:02 +0000 (09:37 -0700)]
Merge pull request #3795 from rzarzynski/wip-10938

rgw: improve support for X-Timestamp header of Swift API

Reviewed-by: Yehuda Sadeh <yehuda@redhat.com>
10 years agoEvent: Let external events ran without lock/unlock 3913/head
Haomai Wang [Mon, 9 Mar 2015 16:26:22 +0000 (00:26 +0800)]
Event: Let external events ran without lock/unlock

Signed-off-by: Haomai Wang <haomaiwang@gmail.com>
10 years agoAsyncConnection: Don't block process when throttle is full
Haomai Wang [Mon, 9 Mar 2015 16:20:24 +0000 (00:20 +0800)]
AsyncConnection: Don't block process when throttle is full

If this connection is blocking on throttler and another worker thread
try to acquire this connection's lock, it will deadlock.

Signed-off-by: Haomai Wang <haomaiwang@gmail.com>
10 years agoMerge pull request #3668 from ceph/wip-compact-container
Gregory Farnum [Mon, 9 Mar 2015 16:17:40 +0000 (12:17 -0400)]
Merge pull request #3668 from ceph/wip-compact-container

Optimize memory usage of MDS

Reviewed-by: John Spray <john.spray@redhat.com>
Reviewed-by: Greg Farnum <gfarnum@redhat.com>
10 years agoMerge pull request #3908 from gravitystorm/patch-2
Loic Dachary [Mon, 9 Mar 2015 15:35:28 +0000 (16:35 +0100)]
Merge pull request #3908 from gravitystorm/patch-2

fix typo

Reviewed-by: Loic Dachary <ldachary@redhat.com>
10 years agodeb: add zlib1g-dev to Build-Depends for Debian/jessie 3909/head
Loic Dachary [Mon, 9 Mar 2015 14:41:21 +0000 (15:41 +0100)]
deb: add zlib1g-dev to Build-Depends for Debian/jessie

The zlib1g-dev is installed indirectly for Ubuntu 12.04 or Ubuntu 14.04
but it is only suggested in Debian/jessie. Adding it to the
Build-depends is redundant and harmless for Ubuntu and resolves the
missing dependency for Debian.

http://tracker.ceph.com/issues/11068 Fixes: #11068

Signed-off-by: Loic Dachary <ldachary@redhat.com>
10 years agodoc: fix typo 3908/head
Andy Allan [Mon, 9 Mar 2015 14:41:18 +0000 (14:41 +0000)]
doc: fix typo

Signed-off-by: Andy Allan <github@gravitystorm.co.uk>
10 years agoMerge pull request #3882 from dachary/wip-install-deps-alternatives
Loic Dachary [Mon, 9 Mar 2015 12:53:25 +0000 (13:53 +0100)]
Merge pull request #3882 from dachary/wip-install-deps-alternatives

install-deps.sh: strip | in the list of packages

Reviewed-by: Reviewed-by: Sage Weil <sage@redhat.com>
10 years agoMerge pull request #3897 from ceph/wip-11015
Sage Weil [Mon, 9 Mar 2015 12:41:21 +0000 (05:41 -0700)]
Merge pull request #3897 from ceph/wip-11015

crush: fix compilation of strwa2 buckets (fixes 11015)

Reviewed-by: Loic Dachary <ldachary@redhat.com>
10 years agoMerge pull request #3899 from yuyuyu101/hammer-test
Sage Weil [Mon, 9 Mar 2015 12:40:49 +0000 (05:40 -0700)]
Merge pull request #3899 from yuyuyu101/hammer-test

TestAsyncDriver: Fix typo in ceph_test_async_driver

10 years agoMerge pull request #3906 from dachary/wip-mailmap
Loic Dachary [Mon, 9 Mar 2015 12:28:17 +0000 (13:28 +0100)]
Merge pull request #3906 from dachary/wip-mailmap

mailmap: David Zhang affiliation

Reviewed-by: Zhi (David) Zhang <zhangz@yahoo-inc.com>
10 years agotests: increase the delay waiting for a cluster event 3907/head
Loic Dachary [Mon, 9 Mar 2015 11:56:31 +0000 (12:56 +0100)]
tests: increase the delay waiting for a cluster event

In some cases the machine running tests may be slow enough that it takes
more than a minute for an OSD to come up. It only happens rarely and
changing the wait period from 60 seconds to 120 seconds. Is so slow
that it takes more than 2 minutes to bring an OSD up, chances are a lot
more will go wrong anyway.

Signed-off-by: Loic Dachary <ldachary@redhat.com>
10 years agomailmap: Zhi (David) Zhang affiliation 3906/head
Loic Dachary [Mon, 9 Mar 2015 10:46:20 +0000 (11:46 +0100)]
mailmap: Zhi (David) Zhang affiliation

And name normalization

Reviewed-by: Zhi (David) Zhang <zhangz@yahoo-inc.com>
Signed-off-by: Loic Dachary <ldachary@redhat.com>
10 years agodoc: preparing a device implies activate 3901/head
Loic Dachary [Sat, 7 Mar 2015 23:18:24 +0000 (00:18 +0100)]
doc: preparing a device implies activate

http://tracker.ceph.com/issues/10922 Fixes: #10922

Signed-off-by: Loic Dachary <ldachary@redhat.com>
10 years agoMerge pull request #3874 from ceph/wip-10828
Sage Weil [Mon, 9 Mar 2015 04:04:19 +0000 (21:04 -0700)]
Merge pull request #3874 from ceph/wip-10828

objecter: fix linger cancel link bug 10828

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
10 years agoMerge pull request #3900 from dachary/wip-recursive-submodules
Loic Dachary [Sat, 7 Mar 2015 17:30:55 +0000 (18:30 +0100)]
Merge pull request #3900 from dachary/wip-recursive-submodules

autogen.sh: do not submodule --force if git does not support it

10 years agoautogen.sh: do not submodule --force if git does not support it 3900/head
Loic Dachary [Sat, 7 Mar 2015 16:11:58 +0000 (17:11 +0100)]
autogen.sh: do not submodule --force if git does not support it

Signed-off-by: Loic Dachary <ldachary@redhat.com>
10 years agoMerge pull request #3895 from dachary/wip-recursive-submodules
Loic Dachary [Sat, 7 Mar 2015 15:53:50 +0000 (16:53 +0100)]
Merge pull request #3895 from dachary/wip-recursive-submodules

doc,tests: force checkout of submodules

10 years agoTestAsyncDriver: Fix typo in ceph_test_async_driver 3899/head
Haomai Wang [Sat, 28 Feb 2015 09:42:52 +0000 (17:42 +0800)]
TestAsyncDriver: Fix typo in ceph_test_async_driver

Fix: #11058
Signed-off-by: Haomai Wang <haomaiwang@gmail.com>
10 years agoReplicatedPG: block write on degraded object if there are waiters 3911/head
Samuel Just [Sat, 7 Mar 2015 02:02:25 +0000 (18:02 -0800)]
ReplicatedPG: block write on degraded object if there are waiters

Suppose we have min_size of 2 and size of 3, foo exists only on the
primary.
- block op 1 on foo due to < min_size
- start recovery on foo for replicas 1 and 2
- complete push to replica 1 (2 copies now)
- allow op 2 on foo through since we have 2 copies
- complete recovery on foo, requeue op 1

Fixes: 11057
Signed-off-by: Samuel Just <sjust@redhat.com>
10 years agocrushtool: test compilation of a rule with straw2 buckets 3897/head
Sage Weil [Sat, 7 Mar 2015 01:47:03 +0000 (17:47 -0800)]
crushtool: test compilation of a rule with straw2 buckets

Signed-off-by: Sage Weil <sage@redhat.com>
10 years agocrush: parse alg names that include a digit
Sage Weil [Sat, 7 Mar 2015 01:43:11 +0000 (17:43 -0800)]
crush: parse alg names that include a digit

The str_p("straw2") won't parse straw2, it seems because the digit is throwing
it off.  Use the existing name rule instead which is more robust.  Note that
not constraining the alg value here is better anyway because instead of a
'cannot parse' error that is hard to debug we instead get a 'unknown alg foo'
error when doing the semantic pass.

Fixes: #11015
Signed-off-by: Sage Weil <sage@redhat.com>
10 years agotests: add Debian jessie dockerfile 3882/head
Loic Dachary [Thu, 5 Mar 2015 11:34:48 +0000 (12:34 +0100)]
tests: add Debian jessie dockerfile

Signed-off-by: Loic Dachary <ldachary@redhat.com>
10 years agotests: DEBIAN_FRONTEND=noninteractive apt-get install
Loic Dachary [Thu, 5 Mar 2015 11:17:26 +0000 (12:17 +0100)]
tests: DEBIAN_FRONTEND=noninteractive apt-get install

Go into non interactive mode when installing the compilation
dependencies, in case a package has a different default mode.

Signed-off-by: Loic Dachary <ldachary@redhat.com>
10 years agotests: jq is not available on Ubuntu precise
Loic Dachary [Thu, 5 Mar 2015 11:06:21 +0000 (12:06 +0100)]
tests: jq is not available on Ubuntu precise

It should be installed from precise-backports but since it is not
currently used, just remove it.

Signed-off-by: Loic Dachary <ldachary@redhat.com>
10 years agoinstall-deps.sh: strip | in the list of packages
Loic Dachary [Thu, 5 Mar 2015 10:38:18 +0000 (11:38 +0100)]
install-deps.sh: strip | in the list of packages

Alternatives were introduced lately and the | needs to be stripped from
the list of packages to install otherwise apt-get will try to install
all packages.

Signed-off-by: Yann Dupont <yann@objoo.org>
10 years agorgw: flush watch after unregistering 3845/head
Yehuda Sadeh [Thu, 5 Mar 2015 21:03:02 +0000 (13:03 -0800)]
rgw: flush watch after unregistering

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
10 years agoMerge pull request #3835 from ceph/wip-mon-datahealth-fix.hammer
Sage Weil [Sat, 7 Mar 2015 00:20:52 +0000 (16:20 -0800)]
Merge pull request #3835 from ceph/wip-mon-datahealth-fix.hammer

[hammer] mon: initialize data store stats structs to avoid weird behaviors

10 years agoMerge pull request #3615 from ceph/wip-perf-watch
John Spray [Fri, 6 Mar 2015 19:18:59 +0000 (19:18 +0000)]
Merge pull request #3615 from ceph/wip-perf-watch

Performance watching CLI

10 years agoMerge pull request #3877 from ceph/hammer-backports-hadoop
Gregory Farnum [Fri, 6 Mar 2015 17:29:22 +0000 (09:29 -0800)]
Merge pull request #3877 from ceph/hammer-backports-hadoop

Hammer backports hadoop

Reviewed-by: Greg Farnum <gfarnum@redhat.com>
10 years agoClient: do not require successful remount when unmounting
Greg Farnum [Tue, 3 Mar 2015 06:29:10 +0000 (22:29 -0800)]
Client: do not require successful remount when unmounting

Fixes: #10982
Signed-off-by: Greg Farnum <gfarnum@redhat.com>
(cherry picked from commit 71f2686252ffe29d0234cc8be7304be1450d090e)

10 years agodoc,tests: force checkout of submodules 3895/head
Loic Dachary [Fri, 6 Mar 2015 16:08:55 +0000 (17:08 +0100)]
doc,tests: force checkout of submodules

When updating submodules, always checkout even if the HEAD is the
desired commit hash (update --force) to avoid the following:

* a directory gmock exists in hammer
* a submodule gmock replaces the directory gmock in master
* checkout master + submodule update : gmock/.git is created
* checkout hammer : the gmock directory still contains the .git from
  master because it did not exist at the time and checkout won't
  remove untracked directories
* checkout master + submodule update : git rev-parse HEAD is
  at the desired commit although the content of the gmock directory
  is from hammer

Signed-off-by: Loic Dachary <ldachary@redhat.com>
10 years agoMerge pull request #3892 from cernceph/cephdisk
Sage Weil [Fri, 6 Mar 2015 15:50:41 +0000 (07:50 -0800)]
Merge pull request #3892 from cernceph/cephdisk

ceph-disk: activate-[all|journal] should suppress

Backport: hammer, giant, firefly
Reviewed-by: Sage Weil <sage@redhat.com>
10 years agotests: sync must also be recursive
Loic Dachary [Fri, 6 Mar 2015 14:29:22 +0000 (15:29 +0100)]
tests: sync must also be recursive

In case a submodule URL to a nested submodule changed

Signed-off-by: Loic Dachary <ldachary@redhat.com>
10 years agoMerge pull request #3718 from ceph/wip-10649
John Spray [Fri, 6 Mar 2015 12:16:09 +0000 (12:16 +0000)]
Merge pull request #3718 from ceph/wip-10649

mds: new OMAP storage for sessionmap

Reviewed-by: Greg Farnum <gfarnum@redhat.com>
10 years agomds: new OMAP storage for sessionmap 3718/head
John Spray [Thu, 12 Feb 2015 10:06:27 +0000 (10:06 +0000)]
mds: new OMAP storage for sessionmap

Fixes: #10649
Signed-off-by: John Spray <john.spray@redhat.com>
10 years agoosdc: add omap op perf counters
John Spray [Thu, 19 Feb 2015 11:01:21 +0000 (11:01 +0000)]
osdc: add omap op perf counters

To avoid bloating the list of counters too much,
group the OMAP operations in to wr,rd,del

Signed-off-by: John Spray <john.spray@redhat.com>
10 years agoceph-disk: activate-[all|journal] should suppress 3892/head
Dan van der Ster [Fri, 6 Mar 2015 10:37:32 +0000 (11:37 +0100)]
ceph-disk: activate-[all|journal] should suppress

Make the suppress-activate <device> feature work for activate-all
and activate-journal.

Signed-off-by: Dan van der Ster <daniel.vanderster@cern.ch>
10 years agoosdc: fix a memory leak in C_TwoContexts 3890/head
Xiong Yiliang [Fri, 6 Mar 2015 04:46:49 +0000 (12:46 +0800)]
osdc: fix a memory leak in C_TwoContexts

If an ObjectOperation op is cancelled, its destructor is
called and each Context object in out_handler is deleted.
A C_TwoContexts object can be one of these handlers. The
two contexts wrapped in C_TwoContexts must be deleted
as well.

Signed-off-by: Xiong Yiliang <xiongyiliang@xunlei.com>
10 years agoMerge pull request #3891 from cernceph/doc
Kefu Chai [Fri, 6 Mar 2015 08:19:10 +0000 (16:19 +0800)]
Merge pull request #3891 from cernceph/doc

doc: osd map cache size is a count, not MB

Reviewed-by: Kefu Chai <kchai@redhat.com>
10 years agodoc: osd map cache size is a count, not MB 3891/head
Dan van der Ster [Fri, 6 Mar 2015 08:05:56 +0000 (09:05 +0100)]
doc: osd map cache size is a count, not MB

Signed-off-by: Dan van der Ster <daniel.vanderster@cern.ch>
10 years agorgw - make starting quota/gc threads configurable 3608/head
Guang Yang [Wed, 4 Feb 2015 03:21:52 +0000 (03:21 +0000)]
rgw - make starting quota/gc threads configurable

Fixes: 11047
Backport: hammer

Signed-off-by: Guang Yang <yguang@yahoo-inc.com>
10 years agoMerge pull request #3888 from dachary/wip-gmock-submodule
Loic Dachary [Thu, 5 Mar 2015 23:55:47 +0000 (00:55 +0100)]
Merge pull request #3888 from dachary/wip-gmock-submodule

ceph needs gmock/gtest to be statically linked

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
10 years agoceph needs gmock/gtest to be statically linked 3888/head
Loic Dachary [Thu, 5 Mar 2015 23:06:21 +0000 (00:06 +0100)]
ceph needs gmock/gtest to be statically linked

The ceph-test package depends on gmock and gtest and needs to statically
link them because packages are not configured to distribute the
corresponding shared library files.

Update the gmock submodule to the version that is configured with
noinst libraries that will be statically linked.

The gmock submodule is modified to reference the branch in which the
ceph specific changes have been done.

http://tracker.ceph.com/issues/11040 Fixes: #11040

Signed-off-by: Loic Dachary <loic@dachary.org>
10 years agoMerge pull request #3880 from jdurgin/wip-librbd-permissions-test
Jason Dillaman [Thu, 5 Mar 2015 21:13:39 +0000 (16:13 -0500)]
Merge pull request #3880 from jdurgin/wip-librbd-permissions-test

librbd: remove bad assert and fix permissions test

Reviewed-by: Jason Dillaman <dillaman@redhat.com>
10 years agorgw: only finialize finisher if it's not null
Yehuda Sadeh [Mon, 16 Feb 2015 18:13:44 +0000 (10:13 -0800)]
rgw: only finialize finisher if it's not null

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
10 years agorgw: fix watch initialization and reinit on error
Yehuda Sadeh [Thu, 12 Feb 2015 23:29:24 +0000 (15:29 -0800)]
rgw: fix watch initialization and reinit on error

We can't just call rados->finalize_watch() and rados->init_watch() from
the watcher, as these calls modify the watcher itself. Also, we can't just
enable the cache after successful watch, because there's more than one
watcher. Only enable the cache if all watchers are set.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
10 years agorgw: move watch reinit into a finisher
Yehuda Sadeh [Thu, 12 Feb 2015 17:56:44 +0000 (09:56 -0800)]
rgw: move watch reinit into a finisher

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
10 years agorgw: add support for new watch/notify functionality
Yehuda Sadeh [Fri, 21 Nov 2014 18:13:29 +0000 (10:13 -0800)]
rgw: add support for new watch/notify functionality

Disable and invalidate cache on watch error, then reinitialize watch,
reenable cache.

Signed-off-by: Yehuda Sadeh <yehuda@redhat.com>
Conflicts:
src/rgw/rgw_cache.h
src/rgw/rgw_rados.cc
src/rgw/rgw_rados.h

10 years agorgw: switch to new watch/notify API
Sage Weil [Wed, 11 Feb 2015 00:20:12 +0000 (16:20 -0800)]
rgw: switch to new watch/notify API

Note that we don't really use it fully, yet.  The main semantic change
here is that we have to explicitly ack the notify.

However, still missing is re-registration of the watch if we see a failure,
and ignoring the cache if watch_check tells us the watch is stale.

Signed-off-by: Sage Weil <sage@redhat.com>
10 years agotest: add unit tests for daemonperf formatting 3615/head
John Spray [Mon, 23 Feb 2015 12:28:32 +0000 (12:28 +0000)]
test: add unit tests for daemonperf formatting

Signed-off-by: John Spray <john.spray@redhat.com>
10 years agoceph.in: add 'daemonperf' command
John Spray [Tue, 3 Feb 2015 13:13:29 +0000 (13:13 +0000)]
ceph.in: add 'daemonperf' command

This is inspired by dstat and scripts/perf-watch.py, to
give a convenient live view of an interesting subset
of the performance counters from a Ceph daemon.

Signed-off-by: John Spray <john.spray@redhat.com>
10 years agomds: fix type of some stats
John Spray [Tue, 3 Feb 2015 17:01:01 +0000 (17:01 +0000)]
mds: fix type of some stats

These guys were marked as counters, but they were
updated with set(), not inc() -- so they should just
be normal u64s.

Signed-off-by: John Spray <john.spray@redhat.com>
10 years agomon: add nicknames to some performance counters
John Spray [Mon, 23 Feb 2015 11:53:02 +0000 (11:53 +0000)]
mon: add nicknames to some performance counters

Signed-off-by: John Spray <john.spray@redhat.com>
10 years agoosd: add nicks to some stats
John Spray [Mon, 23 Feb 2015 11:52:17 +0000 (11:52 +0000)]
osd: add nicks to some stats

This is a rough indication of what I find
useful.

Signed-off-by: John Spray <john.spray@redhat.com>
10 years agoosdc: add perf counter nicknames
John Spray [Wed, 4 Feb 2015 08:14:46 +0000 (08:14 +0000)]
osdc: add perf counter nicknames

Signed-off-by: John Spray <john.spray@redhat.com>
10 years agomds: mark stats with nicknames
John Spray [Tue, 3 Feb 2015 16:48:53 +0000 (16:48 +0000)]
mds: mark stats with nicknames

These are the stats that we would like to appear
in the new perf watch CLI mode.

Signed-off-by: John Spray <john.spray@redhat.com>
10 years agocommon: add optional perf counter nickname
John Spray [Tue, 3 Feb 2015 16:47:32 +0000 (16:47 +0000)]
common: add optional perf counter nickname

This has two purposes:

* Identify counters that are useful in short/summary views
  of a daemon's performance (only these have a nick set)
* Provide shortened versions of names that are suitable
  for display in a text-mode columnar view.

Signed-off-by: John Spray <john.spray@redhat.com>
10 years agocommon: remove some spurious declarations
John Spray [Thu, 5 Mar 2015 15:38:58 +0000 (15:38 +0000)]
common: remove some spurious declarations

Maybe these were defined at some point in history...

Signed-off-by: John Spray <john.spray@redhat.com>
10 years agoMerge pull request #3836 from ceph/wip-10546
Sage Weil [Thu, 5 Mar 2015 20:12:07 +0000 (12:12 -0800)]
Merge pull request #3836 from ceph/wip-10546

mon: Monitor: fix timecheck rounds period

Reviewed-by: Sage Weil <sage@redhat.com>
10 years agoceph.spec.in: loosen ceph-test's dependencies
Ken Dreyer [Wed, 4 Mar 2015 22:01:34 +0000 (15:01 -0700)]
ceph.spec.in: loosen ceph-test's dependencies

In Debian, the ceph-test package can be installed with any version of
ceph-common.

Prior to this commit, in RHEL, we're much more strict about which
version of the dependencies we require. We depend directly on
librados2/librbd1/libcephfs1 instead of ceph-common, and we also require
the specific versions of these libraries to match the version of
ceph-test.

For testing Ceph, it is nice to have the ability to upgrade the
librados2/librbd1/libcephfs1 libraries on a host without having to also
upgrade the ceph-test package as well.

Remove the version number requirements, and change the dependencies from
librados2/librbd1/libcephfs1 to simply "ceph-common". That will make
/etc/ceph/ and /var/log/ceph present for the tests.

http://tracker.ceph.com/issues/10989 Refs: #10989

Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
(cherry picked from commit 924f85f157b8acfe753d944dae49fca390da37aa)

10 years agoMerge pull request #3886 from ceph/wip-gmock-mirror
Loic Dachary [Thu, 5 Mar 2015 16:44:05 +0000 (17:44 +0100)]
Merge pull request #3886 from ceph/wip-gmock-mirror

.gitmodule_mirrors: add gmock.git mirror

Reviewed-by: Loic Dachary <ldachary@redhat.com>
10 years agorgw: improve format of X-Timestamp on Swift objects. 3795/head
Radoslaw Zarzynski [Wed, 25 Feb 2015 15:12:58 +0000 (16:12 +0100)]
rgw: improve format of X-Timestamp on Swift objects.

Backport: hammer
Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>
10 years agorgw: add support for X-Timestamp on Swift containers.
Radoslaw Zarzynski [Wed, 25 Feb 2015 13:11:40 +0000 (14:11 +0100)]
rgw: add support for X-Timestamp on Swift containers.

Fixes: #10938
Backport: hammer
Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>
10 years ago.gitmodule_mirrors: add gmock.git mirror 3886/head
Sage Weil [Thu, 5 Mar 2015 16:29:11 +0000 (08:29 -0800)]
.gitmodule_mirrors: add gmock.git mirror

Signed-off-by: Sage Weil <sage@redhat.com>
10 years agoMerge pull request #3822 from dachary/wip-10947-setcrushmap
Loic Dachary [Thu, 5 Mar 2015 13:28:27 +0000 (14:28 +0100)]
Merge pull request #3822 from dachary/wip-10947-setcrushmap

mon: timeout if crushtool takes longer than mon_lease

Reviewed-by: Joao Eduardo Luis <joao@redhat.com>
10 years agoMerge pull request #3846 from ceph/wip-da-gtest-use-submodules
Kefu Chai [Thu, 5 Mar 2015 10:16:23 +0000 (18:16 +0800)]
Merge pull request #3846 from ceph/wip-da-gtest-use-submodules

Use git submodules for gmock/gtest

Signed-off-by: Kefu Chai <kchai@redhat.com>
10 years agoMerge pull request #3811 from tchaikov/fix-build-debug-xio
Danny Al-Gaaf [Thu, 5 Mar 2015 08:10:48 +0000 (09:10 +0100)]
Merge pull request #3811 from tchaikov/fix-build-debug-xio

Fix build of `--with-debug --enable-xio`

Reviewed-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
10 years agoqa: fix up rbd permissions test 3880/head
Josh Durgin [Thu, 5 Mar 2015 02:38:29 +0000 (18:38 -0800)]
qa: fix up rbd permissions test

! doesn't do what we want in bash -e. Use a more explicit helper
instead, and specify expected error codes.

Signed-off-by: Josh Durgin <jdurgin@redhat.com>
10 years agolibrbd: remove unneeded assert from unregister_watch()
Josh Durgin [Thu, 5 Mar 2015 01:52:39 +0000 (17:52 -0800)]
librbd: remove unneeded assert from unregister_watch()

Calling this unconditionally in close_image() simplifies error
handling when open_image() fails due to an error like EPERM from
registering a watch.

Signed-off-by: Josh Durgin <jdurgin@redhat.com>
10 years agoMerge pull request #3861 from guangyy/wip-11001
Yehuda Sadeh [Thu, 5 Mar 2015 01:01:24 +0000 (17:01 -0800)]
Merge pull request #3861 from guangyy/wip-11001

rgw: do not pre-fetch data for HEAD requests

Reviewed-by: Yehuda Sadeh <yehuda@redhat.com>
10 years agorgw: do not pre-fetch data for HEAD requests 3861/head
Guang Yang [Tue, 3 Mar 2015 09:46:52 +0000 (09:46 +0000)]
rgw: do not pre-fetch data for HEAD requests

Backport: hammer

Signed-off-by: Guang Yang <yguang@yahoo-inc.com>
10 years agoMerge pull request #3876 from ceph/wip-rpm-ceph-test-dependencies
Sage Weil [Thu, 5 Mar 2015 00:11:53 +0000 (16:11 -0800)]
Merge pull request #3876 from ceph/wip-rpm-ceph-test-dependencies

#10989 ceph.spec.in: loosen ceph-test's dependencies

Reviewed-by: Josh Durgin <jdurgin@redhat.com>
10 years agoceph.spec.in: loosen ceph-test's dependencies 3876/head
Ken Dreyer [Wed, 4 Mar 2015 22:01:34 +0000 (15:01 -0700)]
ceph.spec.in: loosen ceph-test's dependencies

In Debian, the ceph-test package can be installed with any version of
ceph-common.

Prior to this commit, in RHEL, we're much more strict about which
version of the dependencies we require. We depend directly on
librados2/librbd1/libcephfs1 instead of ceph-common, and we also require
the specific versions of these libraries to match the version of
ceph-test.

For testing Ceph, it is nice to have the ability to upgrade the
librados2/librbd1/libcephfs1 libraries on a host without having to also
upgrade the ceph-test package as well.

Remove the version number requirements, and change the dependencies from
librados2/librbd1/libcephfs1 to simply "ceph-common". That will make
/etc/ceph/ and /var/log/ceph present for the tests.

http://tracker.ceph.com/issues/10989 Refs: #10989

Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
10 years agoosdc/Objecter: clean up oncommit_sync (and fix leak) 3874/head
Sage Weil [Wed, 4 Mar 2015 19:48:22 +0000 (11:48 -0800)]
osdc/Objecter: clean up oncommit_sync (and fix leak)

This was broken by e0707180dede2dba1ed53f96fd922eda94609480, which doesn't
clean up oncommit_sync for a canceled (linger) request.

Signed-off-by: Sage Weil <sage@redhat.com>
10 years agoosdc/Objecter: count oncommit_sync as uncommitted
Sage Weil [Wed, 4 Mar 2015 18:28:11 +0000 (10:28 -0800)]
osdc/Objecter: count oncommit_sync as uncommitted

Just be consistent.

Signed-off-by: Sage Weil <sage@redhat.com>
10 years agohadoop: workunits don't need java path 3877/head
Noah Watkins [Wed, 25 Feb 2015 02:26:44 +0000 (18:26 -0800)]
hadoop: workunits don't need java path

Signed-off-by: Noah Watkins <noahwatkins@gmail.com>
(cherry picked from commit b3329a961b1ed6bc05969a88743238a552e2b507)

Conflicts:
qa/workunits/hadoop/terasort.sh

10 years agoqa: update old replication tests
Noah Watkins [Tue, 24 Feb 2015 03:59:51 +0000 (19:59 -0800)]
qa: update old replication tests

Signed-off-by: Noah Watkins <noahwatkins@gmail.com>
(cherry picked from commit 15da81008632eaffa83596b034efcca8a51b257e)

10 years agoMerge pull request #3789 from athanatos/wip-9891
Samuel Just [Wed, 4 Mar 2015 15:11:40 +0000 (07:11 -0800)]
Merge pull request #3789 from athanatos/wip-9891

DBObjectMap: lock header_lock on sync()

Reviewed-by: Kefu Chai <kchai@redhat.com>
10 years agoMerge pull request #3804 from Ved-vampir/master
John Spray [Wed, 4 Mar 2015 11:43:25 +0000 (11:43 +0000)]
Merge pull request #3804 from Ved-vampir/master

common: add perf counters description

Reviewed-by: John Spray <john.spray@redhat.com>
Reviewed-by: Loic Dachary <ldachary@redhat.com>
Reviewed-by: Sage Weil <sage@redhat.com>
10 years agoMerge pull request #3869 from tchaikov/print-help-on-osd
Loic Dachary [Wed, 4 Mar 2015 11:26:55 +0000 (12:26 +0100)]
Merge pull request #3869 from tchaikov/print-help-on-osd

ceph.in: print help on 'osd' or 'mon'

Reviewed-by: Loic Dachary <ldachary@redhat.com>
10 years agoceph.in: print help on 'osd' or 'mon' 3869/head
Kefu Chai [Wed, 4 Mar 2015 06:51:34 +0000 (14:51 +0800)]
ceph.in: print help on 'osd' or 'mon'

Fixes: #10888
Signed-off-by: Kefu Chai <kchai@redhat.com>
10 years agocommon: add perf counters description 3804/head
Ved-vampir [Thu, 26 Feb 2015 11:32:29 +0000 (14:32 +0300)]
common: add perf counters description

Signed-off-by: Kiseleva Alyona <akiselyova@mirantis.com>