ceph_manager: do_pg_scrub: keep scrubbing until it's done

The ceph pg scrub ... command isn't really guarranteed to
start a scrub, keep reissuing it until the scrub actually
happens.

Related: #12746
Signed-off-by: Samuel Just <sjust@redhat.com>

Merge pull request #704 from dachary/wip-buildpackages-packaging

buildpackages: protect packages-repository

buildpackages: protect packages-repository

Signed-off-by: Loic Dachary <loic@dachary.org>

Merge pull request #701 from dachary/wip-buildpackages-giant

buildpackages: do not fails for unpackaged files

buildpackages: do not fails for unpackaged files

Older versions of Ceph have that problem failing would break all upgrade suites.

Signed-off-by: Loic Dachary <loic@dachary.org>

Merge pull request #700 from ceph/yuriw-patch-1

Update schedule_rados.sh

Update schedule_rados.sh

Fixed the args order

Merge pull request #699 from ceph/wip_rados_subset

Require -e arg for email address

Require -e arg for email address

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #697 from dachary/wip-buildpackages-giant

buildpackages: giant has some unpackaged files

buildpackages: giant has some unpackaged files

And there really is nothing we can do about it expect ignore them.

Signed-off-by: Loic Dachary <loic@dachary.org>

Merge pull request #695 from dachary/wip-buildpackages-flavor

buildpackages: 10GB is too small most of the time

buildpackages: 10GB is too small most of the time

Signed-off-by: Loic Dachary <loic@dachary.org>

Merge pull request #684 from dachary/wip-12973

rados: 'failed to encode ...' warnings are normal on upgrades

Merge pull request #694 from ceph/wip_fix_f-h-x-offline

Fixing /hammer-x/f-h-x-offline suite

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

Added %
Fixed intall.upgrade syntax
Whitelisted 'reached quota'

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #693 from ceph/wip_fix_f-h-x-offline

Fixed install.upgrade syntax

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

Added %
Fixed intall.upgrade syntax

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #692 from ceph/wip_fix_f-h-x-offline

Added %

Added %

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #689 from ceph/wip_rados_subset

Merge pull request #686 from ceph/wip_rados_subset

Reviewed-by: Samuel Just <sjust@redhat.com>

Merge pull request #686 from ceph/wip_rados_subset

Reducing the ammount of resulting jobs scheduled
Fixed the math added comments

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #688 from ceph/jewel

Reducing the ammount of resulting jobs scheduled

Reducing the ammount of resulting jobs scheduled

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>
(cherry picked from commit 92980c3063487b477c1bff71ebb2c403bcdfa73c)
Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #686 from ceph/wip_rados_subset

Reducing the ammount of resulting jobs scheduled

Reducing the ammount of resulting jobs scheduled

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #681 from ceph/wip-fio-ver

Update fio version from 2.2.9 to 2.2.11

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

rados: 'failed to encode ...' warnings are normal on upgrades

http://tracker.ceph.com/issues/12973 Fixes: #12973

Signed-off-by: Loic Dachary <loic@dachary.org>

Update fio version from 2.2.9 to 2.2.11

Also removes fio-version option from yaml since its redundant and if required can be specified in
overrides

Signed-off-by: Vasu Kulkarni <vasu@redhat.com>

Merge pull request #682 from ceph/wip-13745-jewel

rbd_fio: removed deprecated command-line option to rbd CLI

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

rbd_fio: removed deprecated command-line option to rbd CLI

Fixes: #13745
Signed-off-by: Jason Dillaman <dillaman@redhat.com>

Merge remote-tracking branch 'gh/infernalis'

client-upgrade: hammer-based test should be used on new client

Avoid attempting to test new release features that are not available
on older hammer OSDs.

Fixes: #13304
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
(cherry picked from commit d918ff023e367f997ae27596f28dac3039cb641c)
Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #669 from ceph/wip-13304

client-upgrade: hammer-based test should be used on new client
http://pulpito.ceph.com/teuthology-2015-11-06_08:55:24-upgrade:client-upgrade-hammer-distro-basic-vps/

client-upgrade: hammer-based test should be used on new client

Avoid attempting to test new release features that are not available
on older hammer OSDs.

Fixes: #13304
Signed-off-by: Jason Dillaman <dillaman@redhat.com>

Merge pull request #672 from dachary/wip-buildpackages-ipv6

buildpackages: silently ignore IPv6 addresses

buildpackages: silently ignore IPv6 addresses

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

Merge pull request #652 from ceph/wip-13559

suites/rbd/qemu: excercise CoW support during QEMU testing

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

Merge pull request #639 from ceph/wip-12748-infernalis

rados/singleton-nomsgr/all: add export-after-evict.yaml

Reviewed-by: Samuel Just <sjust@redhat.com>

Merge pull request #667 from ceph/wip-autorepair

tasks/cephfs: simplify test_backtrace_repair

tasks/cephfs: simplify test_backtrace_repair

This was previously using a bunch of files and a small
MDCache limit to force things out of cache.  It is much
simpler to just drop the journal.

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

Merge pull request #670 from ceph/wip-no-async

basic/msgr: remove async and random for now

basic/msgr: remove async and random for now

Signed-off-by: Samuel Just <sjust@redhat.com>

Added $4 for filter out "random.yaml\,async.yaml" tests for now

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>
(cherry picked from commit ef04eaccfa2e11286d68fb84e9dc66a0ff2a430b)

Merge pull request #666 from ceph/wip-openstack-fixes

Added $4 for filter out "random.yaml\,async.yaml" tests for now

Added $4 for filter out "random.yaml\,async.yaml" tests for now

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #656 from ceph/wip-frag

fs: fix two frag_enable fragments

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

Merge pull request #658 from dachary/wip-buildpackages-tags

buildpackages: fetch tags from the official Ceph repository

buildpackages: fetch tags from the official Ceph repository

A clone of Ceph is not automagically updated with the tags from the
official Ceph repository. For a pull request based on master, git
describe will use whatever tags existed at the time the clone was made,
unless the author pull them from the official Ceph repository and later
git push --tags them.

The output of git describe is used to name the packages and if the
official tags are not present, the packages will be incorrectly
named. For instance instead of 9.0.3-34 the packages could be named
0.87-8433 because the v0.87 tag is the most recent tag in the
repository. That confuses the install task that will fail with:

   'ceph version 0.87 was not installed, found 9.0.3.'

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

Merge pull request #655 from dachary/wip-13450-openstack-rados

openstack: rados/thrash: allocate three disks, always

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

fs: fix two frag_enable fragments

Badly formatted yaml

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

openstack: rados/thrash: allocate three disks, always

The thrasher needs disk attached to run against xfs, ext4 or btrfs. And
some jobs use more disks and do not fit in 40GB.

http://tracker.ceph.com/issues/13450 Fixes: #13450

Signed-off-by: Loic Dachary <loic@dachary.org>

Merge pull request #654 from dachary/wip-ceph-disk

ceph-disk: switch to using install / ceph

openstack: resource hint must contain an array

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

suites/rbd/qemu: excercise CoW support during QEMU testing

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

ceph-disk: switch to using install / ceph

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

ceph: log which ceph.conf file is written

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

Merge pull request #622 from dachary/wip-13451-openstack-rados-objectstore

rados/objectstore/alloc-hint: need three attached disks

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

Merge pull request #645 from ceph/wip-openstack-fixes

Added openstack config openstack.yaml

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

Merge remote-tracking branch 'gh/infernalis'

Added openstack config openstack.yaml
Added +

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #648 from ceph/wip-f-h-x-offline

test offline upgrade from firefly -> x

Reviewed-by: David Zafman <dzafman@redhat.com>

upgrade/hammer-x/f-h-x-offline: test firefly -> x upgrade

This is an offline (osds all restart) upgrade.

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

Merge pull request #644 from ceph/wip-fix-f-h-x-branch

Added one more "write_append_excl: false"

Added one more "write_append_excl: false"

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Added "write_append_excl: false"

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>
(cherry picked from commit 3a13ec04ce68bbc152fcde2a31c573bb6b454219)
Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #643 from ceph/wip-fix-f-h-x-branch

Added "write_append_excl: false"

Added "write_append_excl: false"

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Removed osd.6 from roles to fix #13515

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>
(cherry picked from commit 8ae31f15836d84dc886224ff8849b9bb3dd07e81)
Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #642 from ceph/wip-13515

Removed osd.6 from roles to fix #13515

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

Removed osd.6 from roles to fix #13515

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #641 from ceph/wip-13485

Added write_append_excl: false to fix #13485

Added write_append_excl: false to fix #13485

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Removed /parallel/7-final-workload/rgw_swift.yaml as it was creating the same user as the s3tests run right before

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>
(cherry picked from commit 64dcc28e14f43e68c9253621c4ae7479f9673381)
Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

Merge pull request #640 from ceph/wip-fix-f-h-x-branch

Removed /parallel/7-final-workload/rgw_swift.yaml as it was creating …

Removed /parallel/7-final-workload/rgw_swift.yaml as it was creating the same user as the s3tests run right before

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>

rados/singleton-nomsgr/all: add export-after-evict.yaml

add a test for http://tracker.ceph.com/issues/12748

Fixes: #12748
Signed-off-by: Kefu Chai <kchai@redhat.com>

Merge pull request #632 from ceph/wip-rbd-dir-fix

use cephtest as base dir during testing

Use cephtest as base dir for testing and let nuke cleanup in case of abrupt failures

Signed-off-by: Vasu Kulkarni <vasu@redhat.com>

rgw: sleep before agent startup

For good measure.. it's racy.

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

rgw/verify: allow rgw's to start up before starting agent

Startup is slow(er) under valgrind.

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

rgw/multifs: run multipart (perl) test on ubuntu

no Net::Amazon::S3 on el7

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

Merge remote-tracking branch 'gh/infernalis'

Merge pull request #635 from ceph/wip-cpan

rgw: do quota tests on ubuntu

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

rgw: do quota tests on ubuntu

The Amazon/S3.pm perl module from CPAN is not available as an RPM.

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

Merge pull request #630 from dachary/wip-13408-stress-split

upgrade/stress-split: needs only six OSDs

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

Merge pull request #623 from dachary/wip-11260-openstack-singleton-nomsgr

rados/singleton-nomsgr/all/msgr: needs 15GB RAM

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

tasks/cephfs: fix TestClientLimits.test_client_oldest_tid

To get the health warning, first we need to make sure requests are
added to session's completed request list. Then we need to send an
extra request to MDS to trigger the code that generates the warning.

Fixes: #13437
Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit b03ac0798058e543a80aaab88101314ab63dd22b)

Merge pull request #624 from ceph/wip-13437

Wip 13437 (tasks/cephfs: fix TestClientLimits.test_client_oldest_tid …)

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

upgrade/stress-split: needs only six OSDs

The large number of OSDs is only needed for LRC or SHEC in some
cases. All other jobs will do fine with six OSDs.

http://tracker.ceph.com/issues/13408 Fixes: #13408

Signed-off-by: Loic Dachary <loic@dachary.org>

Merge pull request #628 from dachary/wip-13454-openstack-flavor

buildpackages: implement notcmalloc flavor

buildpackages: implement notcmalloc flavor

http://tracker.ceph.com/issues/13454 Fixes: #13454

Signed-off-by: Loic Dachary <loic@dachary.org>

Merge pull request #621 from dachary/wip-13449-admin-socket

tasks/admin_socket.py: wait 120 seconds instead of 60

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

Merge pull request #619 from dachary/wip-buildpackages-submodules

buildpackages: get ceph submodules

tasks/cephfs: fix TestClientLimits.test_client_oldest_tid

To get the health warning, first we need to make sure requests are
added to session's completed request list. Then we need to send an
extra request to MDS to trigger the code that generates the warning.

Fixes: #13437
Signed-off-by: Yan, Zheng <zyan@redhat.com>

rados/singleton-nomsgr/all/msgr: needs 15GB RAM

If given only 8GB RAM,  ceph_test_msgr may abort with buffer::bad_alloc.

http://tracker.ceph.com/issues/11260 Fixes: #11260

Signed-off-by: Loic Dachary <loic@dachary.org>

rados/objectstore/alloc-hint: need three attached disks

So they can be mkfs xfs.

http://tracker.ceph.com/issues/13451 Fixes: #13451

Signed-off-by: Loic Dachary <loic@dachary.org>

tasks/admin_socket.py: wait 120 seconds instead of 60

When running on virtual machines, it may take more than one minute for a
daemon to create the admin socket.

http://tracker.ceph.com/issues/13449 Fixes: #13449

Signed-off-by: Loic Dachary <loic@dachary.org>

Merge remote-tracking branch 'gh/infernalis'

Merge pull request #618 from yehudasa/wip-infernalis-fix

radosgw_admin: skip log objects that start with obj_delete_at_hint