os/bluestore: use typedef for blob unused (uint16_t)

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

os/bluestore: fsck: verify blob.unused field

Two checks:

- verify that no logical extents reference the portion of the
  blob marked unused, and
- verify that the csum (if present) is zero for any unused
  region.

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

os/bluestore: do not extend blobs with unused bitmap

If we resize the blob we need to adjust the resolution of the
unused bitmap, and that is only possible for some bit patterns.
For now just ignore blobs with unused blocks.

Add an assert in add_tail() so that we don't forget that
add_tail is (probably) where we'd would do that adjustment.

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

os/bluestore: some whitespace

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

os/bluestore: mark_used for wctx->writes

The refactored blob reuse code needs to mark blobs used when
processing wctx->writes.

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

os/bluestore: prefix fsck errors with 'fsck error:'

Greppable!

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

Merge pull request #14275 from ceph/evelu-cephdisk-race

ceph-disk: Adding retry loop in get_partition_dev()

Reviewed-by: Loic Dachary <ldachary@redhat.com>

Merge pull request #14185 from DHB-liuhong/objects

mon/PGMap: make si units more readable in PGMap summary

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

Merge pull request #14313 from wjin/clean_mgr

mgr: rm nonused main function

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

ceph-disk: Adding retry loop in get_partition_dev()

There is very rare cases where get_partition_dev() is called before the actual partition is available in /sys/block/<device>.

It appear that waiting a very short is usually enough to get the partition beein populated.

Analysis:
update_partition() is supposed to be enough to avoid any racing between events sent by parted/sgdisk/partprobe and
the actual creation on the /sys/block/<device>/* entrypoint.
On our CI that race occurs pretty often but trying to reproduce it locally never been possible.

This patch is almost a workaround rather than a fix to the real problem.
It offer retrying after a very short to be make a chance the device to appear.
This approach have been succesful on the CI.

Note his patch is not changing the timing when the device is perfectly created on time and just differ by a 1/5th up to 2 seconds when the bug occurs.

A typical output from the build running on a CI with that code.
command_check_call: Running command: /usr/bin/udevadm settle --timeout=600
get_dm_uuid: get_dm_uuid /dev/sda uuid path is /sys/dev/block/8:0/dm/uuid
get_partition_dev: Try 1/10 : partition 2 for /dev/sda does not in /sys/block/sda
get_partition_dev: Found partition 2 for /dev/sda after 1 tries
        get_dm_uuid: get_dm_uuid /dev/sda uuid path is /sys/dev/block/8:0/dm/uuid
get_dm_uuid: get_dm_uuid /dev/sda2 uuid path is /sys/dev/block/8:2/dm/uuid

fixes: #19428

Signed-off-by: Erwan Velu <erwan@redhat.com>

mgr: rm nonused main function

Signed-off-by: Wei Jin <wjin.cn@gmail.com>

Merge pull request #14174 from shinobu-x/wp-sk-typo-osdmonitor-20170328

mon/OSDMonitor: Clean up: delete extra S signature for plural

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

Merge pull request #12243 from rzarzynski/wip-rgw-bulkupload

rgw: add support for the BulkUpload of Swift API
Reviewed-by: Orit Wasserman <owasserm@redhat.com>

Merge pull request #14120 from wangzhengyong/uninitialized

rgw: fix uninitialized fields

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

Merge pull request #14217 from Liuchang0812/wip-18584-pg-info-dump

osd: print pg_info_t::purged_snaps as array, not string

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

Merge pull request #14036 from batrick/mds-debug-reply

mds: bump client_reply debug to match client_req

Reviewed-by: Yan, Zheng <zyan@redhat.com>

Merge pull request #14296 from theanalyst/fix/script/credits

script: drop the -x arg for credits script

Reviewed-by: Loic Dachary <ldachary@redhat.com>

Merge pull request #13401 from liewegas/wip-kill-subop

osd: kill all remaining MOSDSubOp users

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

Merge pull request #14221 from jcsp/wip-19253-b

qa: whitelist another fullness log message

Reviewed-by: Brad Hubbard <bhubbard@redhat.com>

Merge pull request #14279 from jimmyway/wip-fix-pg-comment

osd: fix comments about pg refs and lock

Reviewed-by: Pan Liu <liupan1111@gmail.com>

Merge pull request #14277 from atheism/fix-test-objectstore-chain_xattr

test: objectstore: chain_xattr: fix wrong memset usage to fill buf

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

Merge pull request #13800 from xiexingguo/wip-clean-pgcreate-throttle

common/config_opts.h: kill mon_pg_create_interval

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

Merge pull request #13621 from Jing-Scott/add-remov-x-delete-feature

rgw: add the remove-x-delete feature to cancel swift object expiration

Reviewed-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

Merge pull request #13862 from jcsp/wip-16523

qa, mds: add checks for fragmentation, and enable it by default

mon/OSDMonitor: Clean up: delete extra S signature for plural

Signed-off-by: Shinobu Kinjo <shinobu@redhat.com>

Merge pull request #14156 from smithfarm/release-notes-add-rbd-mirror

tools: ceph-release-notes: prefixes and pep8 compliance

Reviewed-by: Loic Dachary <ldachary@redhat.com>
Reviewed-by: Abhishek Lekshmanan <abhishek@suse.com>

script: drop the -x arg for credits script

while useful for debugging creates a whole lot of output which we don't
need normally

Signed-off-by: Abhishek Lekshmanan <abhishek@suse.com>

Merge pull request #12084 from hansbogert/patch-4

rgw: Add --zonegroup-new-name  in usage
Reviewed-by: Orit Wasserman <owasserm@redhat.com>

Merge pull request #14269 from liewegas/wip-no-spinlock

mon/OSDMonitor: spinlock -> std::mutex

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

Merge pull request #14274 from liewegas/wip-osd-beacon-fix

osd: fix some osd beacon bugs

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

Merge pull request #14288 from Adirl/small_leaks

msg/async/rdma: Fix small memory leaks detected by valgrind

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

rgw: add the check for Content-Length in BulkUpload.

Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

rgw: implement the full response generation in BulkUpload of Swift API.

Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

rgw: optimize metadata caching in BulkUpload of Swift API.

Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

rgw: implement the object creation in BulkUpload of Swift API.

Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

rgw: implement the container creation in BulkUpload of Swift API.

Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

msg/async/rdma: Fix small memory leaks detected by valgrind

Issue: 995322
Change-Id: Iccffbd0defd09aa51c4b51b78733921444ea1c5f
Signed-off-by: Amir Vadai <amir@vadai.me>

rgw: implement the AlignedStreamGetter.

Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

rgw: implement the basic security check for BulkUpload of Swift API.

Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

Merge pull request #14170 from liewegas/wip-fewer-warnings

common/ceph_context: fewer warnings about experimental features

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

Merge pull request #14271 from joscollin/wip-krbd-warning

rbd: warning, ‘devno’ may be used uninitialized in this function

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

Merge pull request #13909 from kestrels/wip-19217

crc32c: Add crc32c function optimized for ppc architecture

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

Merge pull request #14124 from Fabian-Gruenbichler/ipv6_fix

common: fix segfault in public IPv6 addr picking

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

Merge pull request #14173 from badone/wip-do_request-abort

common: Fix heap buffer overflow in do_request

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

Merge pull request #14193 from liupan1111/wip-fix-delete-full

librados: set the flag CEPH_OSD_FLAG_FULL_TRY of Op in the right place.

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

Merge pull request #14248 from badone/wip-simple_spin_lock-delay

common: Make spinlock delay more conventional

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

Merge pull request #13690 from jcsp/wip-17503

mgr: pass through cluster log to plugins

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

rgw: implement the TAR extraction loop of Swift's BulkUpload.

Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

rgw: implement the TAR format interpreter.

Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

rgw: implement RGWBulkUploadOp::DecoratedStreamGetter.

Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

rgw: initial class structure for BulkUpload of Swift API.

Signed-off-by: Radoslaw Zarzynski <rzarzynski@mirantis.com>

osd: fix comments about pg refs and lock

Signed-off-by: tang.jin <tang.jin@istuary.com>

Merge pull request #14276 from atheism/rgw-rgwop-remove-unused-iter

rgw: rgw_op: remove unused variable iter
Reviewed-by: Orit Wasserman <owasserm@redhat.com>

test: objectstore: chain_xattr: fix wrong memset usage to fill buf

    Current usage:
        memset(buf, sizeof(buf), 0x1F);
    should be:
        memset(buf, 0x1F, sizeof(buf));
    to fill the buffer with '0x1F'.

Signed-off-by: Weibing Zhang <zhangweibing@unitedstack.com>

rgw: rgw_op: remove unused variable iter

    The variable iter is not used in  RGWInitMultipart::execute().

Signed-off-by: Weibing Zhang <zhangweibing@unitedstack.com>

Merge pull request #13811 from rzarzynski/wip-bs-bitmap-mem-cleanup

os/bluestore: memory and dereference clean-up in the BitAllocator

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

Merge pull request #14121 from majianpeng/bluefs-get-allocated

os/bluestore/BlueFS: optimize get_allocated

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

Merge pull request #14162 from majianpeng/optimize-bluestore-force-flush

os/bluestore: clean up flush logic

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

Merge pull request #14225 from ivancich/wip_fix_throttle_config_change

os/bluestore: make live changes for BlueStore throttle config work like initial config

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

osd: only send beacon on connect if active

Otherwise the mon will just ignore it and we'll wait another
cycle.

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

osd: fix beacon condition

- && should have been ||
- change this to instead not send when monmap epoch==0,
  and only set last_sent_beacon if we actually send.  That
  way we'll send on the next tick after we get a real monmap.

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

osd/ReplicatedBackend: remove MOSDSubOp cruft from repop_commit

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

osd: drop unused MOSDSubOp includes

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

osd/ReplicatedBackend: rename sub_op_modify* do_repop*

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

osd/ReplicatedBackend: no empty MOSDSubOps anymore

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

osd/ReplicatedBackend: use more __func__

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

osd: drop unnecessary declarations in headers

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

osd: use MOSDScrubReserve instead of MOSDSubOp for scrub reservations

This is the last MOSDSubOp user.

Note that while the next step is to move to AsyncReserver internally,
this isn't quite yet possible since AsyncReserve "blocks" indefinitely
so we wouldn't generate a REJECT.  Changing how we schdule scrubs
internally will take a bit more work.

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

osd: use new MOSDRepScrubMap message for ScrubMap instead of MOSDSubOp

This is one of the last remaining users of MOSDSubOp!

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

osd: use new MOSDPGBackfillRemove instead of MOSDSubOp

One of the last remaining MOSDSubOp users!

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

osd/ReplicatedBackend: remove legacy push/pull op handling

These are never sent, and thus never recieved.

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

Merge pull request #14229 from ceph/wip-systemd

qa: Add reboot case for systemd test

Merge pull request #14254 from idryomov/wip-vstart-runner-ps

qa/vstart_runner: amend ps invocation

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

Merge pull request #14226 from ceph/preset-mon

systemd: remove ceph-create-keys from presets

Merge pull request #14266 from tchaikov/wip-require-mgr-for-pg-dump

tests: add MGR=1 so 'pg dump' won't be blocked

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

Merge pull request #14259 from johnwilkins/wip-doc-multisite

doc: wip-doc-multisite ports downstream multisite document upstream.

rbd: warning, ‘devno’ may be used uninitialized in this function

The following warning appears during make. Fixed in both unmap_image() functions in krbd.cc

--
krbd.cc: In function ‘int krbd_unmap_by_spec(krbd_ctx*, const char*, const char*, const char*, const char*)’:
krbd.cc:608:65: warning: ‘devno’ may be used uninitialized in this function [-Wmaybe-uninitialized]
   return do_unmap(ctx->udev, devno, build_unmap_buf(id, options));
                                                                 ^
krbd.cc:591:9: note: ‘devno’ was declared here
   dev_t devno;
--

Signed-off-by: Jos Collin <jcollin@redhat.com>

mon/OSDMonitor: spinlock -> std::mutex

I think spinlock is dangerous here: we're doing semi-unbounded
work (decode).  Also seemingly innocuous code like dout macros
take mutexes.

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

osd/OSDMap: use _pg_to_up_acting_osds() to get_pg_acting_primary()

avoid unnecessary swap() call of vector<int> which will be dropped on
the floor after getting its first element, which can be retrieved using
another parameter using _pg_to_up_acting_osds() directly.

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

mgr: use OSDMap::get_pg_acting_primary() to get the primary

OSDMap::pg_to_acting_osds() assumes that the "acting" parameter is not
null.

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

tests: add MGR=1 so 'pg dump' won't be blocked

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

common/config_opts: kill mon_pg_create_interval

This is not needed anymore.

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

Merge pull request #14260 from dmick/wip-workunit-clone-with-branch

workunit: request branch when cloning

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

common: Make spinlock delay more conventional

The accepted method of implementing a spinlock delay is the "pause"
instruction.

Signed-off-by: Brad Hubbard <bhubbard@redhat.com>

tasks/workunit.py: specify the branch name when cloning a branch

c1309fb failed to specify a branch when cloning using --depth=1, which
by default clones the HEAD. and we can not "git checkout" a specific
sha1 if it is not HEAD, after cloning using '--depth=1', so in this
change, we dispatch "tag", "branch", "HEAD" using three Refspec classes.

Signed-off-by: Kefu Chai <kchai@redhat.com>
Signed-off-by: Dan Mick <dan.mick@redhat.com>

rgw: fix uninitialized fields

Signed-off-by: wangzhengyong@cmss.chinamobile.com

Merge pull request #14219 from ShiqiCooperation/master

os/bluestore: clean up Invalid return value judgment

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

Merge pull request #14228 from ifed01/wip-bluestore-fix-store-test

test/store_test: fix bluestore test cases disablement

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

Merge pull request #13617 from liewegas/wip-mgr-commands

mon,mgr: tag some commands for ceph-mgr

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

mon/ConfigKeyService: wait for quorum

Among other things, this prevents a mgr activation stall if it
happens to request config-keys from a mon that is out of
quorum.

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

mon/MgrMonitor: show delta

This is mostly because I couldn't debug a weird state where
it kept updating but it appeared the addr was not changing
(always '-').

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

mon/MgrMap: implement dump, add to ceph-dencoder

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

mon/MgrMonitor: print MgrMap to log on each change

The other services do this.

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

mon/MgrMonitor: health warn/err if no active mgr

Start warning once mons are luminous; start erroring once
require_luminous is set in osdmap.  Allow a grace period for
mgr to restart or standby to take over before we turn a warning
into an error.

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

Ported the downstream (and tested) document upstream.

Signed-off-by: John Wilkins <jowilkin@redhat.com>

Clarified index.

Signed-off-by: John Wilkins <jowilkin@redhat.com>

doc: Added a new version of the multi-site graphic.

Signed-off-by: John Wilkins <jowilkin@redhat.com>

Merge pull request #14183 from joscollin/wip-rgw-cleanup-removeunwantedheaders

rgw: Removed Unwanted headers
Reviewed-by: Orit Wasserman <owasserm@redhat.com>

Merge pull request #14194 from joscollin/wip-rgw_rados-warning

rgw: warning, output may be truncated before the last format character

Reviewed-by: Brad Hubbard <bhubbard@redhat.com>

qa/vstart_runner: amend ps invocation

"ps -xwwu<id>" is parsed as BSD, because -x is not a UNIX option.
"u" is a BSD option for user-oriented format, so the <id> ends up being
parsed as an old-style "select by pid".  The only reason this command
doesn't dump other user's processes is that the BSD "only yourself"
restriction is in effect.

I'm not sure what's wrong with a simple "ps xww", but if we want to
select by euid, let's do it right.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>