Merge pull request #526 from ceph/wip-stashed-layout

tasks/cephfs: implement TestDataScan.test_stashed_layout

Merge pull request #536 from ceph/wip-11881-multipath

ceph-deploy: hints for OpenStack

update the list of supported distributions

CentOS 6 and Ubuntu 12.04 are not support after Infernalis.
CentOS 7 is supported starting with Infernalis.

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

Merge pull request #537 from dachary/wip-supported

update the list of supported distributions

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

update the list of supported distributions

CentOS 6 and Ubuntu 12.04 are not support after Infernalis.
CentOS 7 is supported starting with Infernalis.

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

ceph-deploy: hints for OpenStack

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

Merge pull request #531 from ceph/wip-mds-debug

tasks/kcephfs: enable MDS debug

Merge pull request #533 from ceph/revert-525-wip-12625-master

Revert "Added singleton hammer-x test to address #12625"

Revert "Added singleton hammer-x test to address #12625"

Merge pull request #525 from ceph/wip-12625-master

Added singleton hammer-x test to address #12625

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

Merge pull request #514 from ceph/wip-cd-add-branch-install

 - Use branch from config for ceph-deploy testing

Reviewed-by: Travis Rhoden <trhoden@redhat.com>

Merge pull request #529 from ceph/wip-calamari-cli

calamari_setup: install "cli" utils on Calamari node

Reviewed-by: Dan Mick <dmick@redhat.com>

tasks/kcephfs: enable MDS debug

To help us debug #11482

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

tasks/cephfs: implement TestDataScan.test_stashed_layout

This is for testing how cephfs-data-scan handles
the new 'layout' xattr when found during recovery.

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

calamari_setup: install "cli" utils on Calamari node

In our RHCS 1.3 ceph-deploy docs, we tell users to run "ceph-deploy
install --cli" on their calamari admin node, but our smoke test wasn't
actually doing this.

See https://bugzilla.redhat.com/1252929 , "[Ubuntu 1.3.0] - ceph-deploy
install --no-adjust-repos --cli `hostname` is failing with a Traceback
error"

Merge pull request #527 from dmick/master

calamari_setup: change to use ceph-deploy repo command

Reviewed-by: Ken Dreyer <kdreyer@redhat.com>

calamari_setup: change to use ceph-deploy repo command

Fixes: #12643
Signed-off-by: Dan Mick <dan.mick@redhat.com>

Merge pull request #524 from ceph/wip-kdreyer-calamari-1.2

calamari_setup: handle RHCS 1.2 (missing /mnt/MON)

Reviewed-by: Dan Mick <dmick@redhat.com>

Added singleton hammer-x test to address #12625

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

Merge pull request #445 from dachary/wip-11666-erasure-code-lrc

erasure-code: lrc plugin workload

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

calamari_setup: handle RHCS 1.2 (missing /mnt/MON)

In RHCS 1.2 we don't have a /mnt/MON directory. The intention of
35c6363a1eaf5011e26be8561ffcd4e9324d13a2 was to handle this condition,
but in 1.2, the non-zero return code makes Teuthology fail the whole
test.

We don't want *Teuthology* itself to act on the return code here; we
simply want to know what it was and structure the rest of the test
accordingly.

Merge pull request #522 from ceph/syntax-check

Include `SyntaxError` checks in linter

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

fix indentation error in docstring

Signed-off-by: Alfredo Deza <adeza@redhat.com>

do not ommit syntax errors in linter checks

Signed-off-by: Alfredo Deza <adeza@redhat.com>

Merge pull request #521 from ceph/wip-fix-ceph

tasks: fix syntax error in ceph.py

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

tasks: fix syntax error in ceph.py

From e195f9fa.

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

Merge pull request #520 from ceph/wip-logrotate-chcon

Set the SELinux context of the logrotate config

Merge pull request #519 from ceph/wip-silent-siganl

tasks/ceph_manager: be silent about sending SIGHUPs

Set the SELinux context of the logrotate config

Signed-off-by: Zack Cerza <zack@redhat.com>

Merge pull request #518 from ceph/fix-async-test

Fix experimental feature override makes user config useless

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

tasks/ceph_manager: be silent about sending SIGHUPs

At the default interval this generates tons of log noise.

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

Fix experimental feature override makes user config useless

Signed-off-by: Haomai Wang <haomaiwang@gmail.com>

Merge pull request #515 from ceph/wip-12476

thrashosds: send a random osd a signal.SIGHUP

Merge pull request #510 from dmick/master

calamari_setup: Add --no-adjust-repos to ceph-deploy install --<type>

Reviewed-by: Travis Rhoden <trhoden@redhat.com>

calamari_setup: only stop ceph on osd/mon machines

Fixes bogus failure at end of run on calamari server

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

calamari_setup: hack for 1.3.0 on centos: install lttng from EPEL

lttng is not yet part of any private repo; since 1.3.0/CentOS is not
a supported product, just grab it from EPEL for this test.

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

calamari_setup: pep8 fix

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

calamari_setup: Add --no-adjust-repos to ceph-deploy install --<type>

Otherwise, ceph-deploy will install an apt source that points to
ceph.com, which will override the local ISO repos.

No --mon/--osd yet until 12147 is fixed

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

Merge pull request #446 from ceph/wip-krbd-unmap-suite

unmap subsuite for krbd

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

ceph_manager: don't add an osd to live_osds until it's been revived

also waits to remove it from dead_osds. this fixes an issue where
do_sighup tries to send a signal to an osd that has not been revived
yet.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

thrashosds: adds a sighup_delay option, defaulted to 0.1

This will call Thrasher.do_sighup which picks a random osd and sends a
signal.SIGHUP to it, delaying for the value of sighup_delay between each
time it picks a new osd to signal.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

ceph_manager: adds a do_sighup method

This method runs in a separate greenlet than do_thrash and will pick a
random live osd to send a signal.SIGHUP to. There is a config option,
sighup_delay, which controls how long to delay between sending the
signals.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

ceph_manager: adds a signal_osd method

Signed-off-by: Andrew Schoen <aschoen@redhat.com>

 - Use branch from config for ceph-deploy testing instead of released branch

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

Merge pull request #509 from ceph/wip-12225

Don't pass host to mon create-initial

Reviewed-by: Travis Rhoden <trhoden@redhat.com>

Merge pull request #512 from ceph/wip-pool-perm-fixes

Fixes to test_pool_perm

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

Merge pull request #489 from dachary/wip-12320-ceph-fuse

tasks/ceph_fuse.py: virtual machines need flexible mount timeout

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

tasks/cephfs: reset auth caps in setUp

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

tasks/cephfs: don't hardcode data pool name in test_pool_perm

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

tasks/cephfs: test_pool_perm: reinstate proper caps on teardown

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

tasks/cephfs: don't hardcode client ID in test_pool_perm

Our teuthology yamls all happen to use "client.0" but
it's not guaranteed.

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

tasks/cephfs: test_pool_perm whitespace

PEP8-ize.

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

Merge pull request #511 from ceph/wip-drop-chef

Remove chef task

Remove chef task

It is obsoleted by teuthology.task.ansible.cephlab

Signed-off-by: Zack Cerza <zack@redhat.com>

 - Fix for wip-12225, Extra argument to mon create-initial was never required

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

Merge pull request #508 from ceph/wip-12460

Fixes #508
Reviewed-by: Samuel Just <sjust@redhat.com>

Fixes #12460

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

Merge pull request #506 from ceph/wip-12068-master

Fixes #506
Reviewed-by: Samuel Just <sjust@redhat.com>

Merge pull request #505 from dmick/master

calamari_setup: disable epel again after ceph-deploy

Reviewed-By: Travis Rhoden <trhoden@redhat.com>

Fixes #12068

point-to-point test moved from upgrade/hammer and modified to run as -x type upgrade

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

calamari_setup: disable epel again after ceph-deploy

ceph-deploy enables EPEL on Centos; this caused problems for
Calamari.  Undo it after ceph-deploy has installed Ceph.

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

Merge pull request #500 from ceph/wip-12425-master

Fixes #12425, moved versions-steps-x suite from upgrade/firefly to up…

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

Fixes #12425, moved versions-steps-x suite from upgrade/firefly to upgrade/firefly-x
as the old location was wrong and never ran correctely to upgrade tp -x

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

Merge pull request #496 from ceph/wip-hammer-x

Added config settings to resolve and test #496

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

Added config settings to resolve and test #12021
"mon debug unsafe allow tier with nonempty snaps: true"
(only for firefly-x, giant-x and upgrade/hammer/point-to-point running againts hammer branch)

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

Merge remote-tracking branch 'gh/next'

Drop chef from the teuthology suite

Chef has been replaced by ansible, which will be added to jobs at
schedule time.

Signed-off-by: Zack Cerza <zack@redhat.com>

Drop chef from tgt suite

Chef has been replaced by ansible, which will be added to jobs at
schedule time.

Signed-off-by: Zack Cerza <zack@redhat.com>
(cherry picked from commit 5b57b81e90393d42494c7b02fb263f1601421f16)

Merge pull request #492 from ceph/wip-drop-chef

Drop chef from suites

Drop chef from tgt suite

Chef has been replaced by ansible, which will be added to jobs at
schedule time.

Signed-off-by: Zack Cerza <zack@redhat.com>

Drop chef from the teuthology suite

Chef has been replaced by ansible, which will be added to jobs at
schedule time.

Signed-off-by: Zack Cerza <zack@redhat.com>

Merge pull request #490 from ceph/wip-12029

suites/rbd: add EC pool variant for QEMU tests

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

Merge pull request #487 from dmick/master

calamari_setup: handle new structure in 1.3.0 (MON and OSD repos)

tasks/ceph_fuse.py: virtual machines need flexible mount timeout

1) add a wait time before the mount attempt to let the cluster get set up.
By default this should be skipped, but for VMs and known-slow systems we
can give them 60 seconds.
2) Make the timeout configurable, with a 30-second default, but override it
for VM tests.

http://tracker.ceph.com/issues/12320 Fixes: #12320

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

Merge pull request #464 from ceph/wip-offline-backward

Tests for damage handling and cephfs-data-scan

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

calamari_setup: handle new structure in 1.3.0 (MON and OSD repos)

Also, clean up the command handling; failure will throw an exception,
so no need to accumulate the errors in an int

Fixes: #12228
Signed-off-by: Dan Mick <dan.mick@redhat.com>

Merge pull request #465 from ceph/wip-11570

hammer suite split - older and newer

Merge pull request #384 from ceph/wip-dirfrag-enable

suites/fs: enable directory fragmentation

suites/fs: enable directory fragmentation

Setting this in tests globally as a precursor
to enabling it by default in the shipping code.
Set a low mds bal split size in order to make
us hit the fragmentation code more often in
than we usually would in a single MDS environment.

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

Merge pull request #470 from ceph/wip-4161

Add cephfs TestBacktrace

Merge pull request #466 from ceph/wip-damaged-fixes

Tests for wip-damaged-fixes

Merge pull request #485 from athanatos/wip-12231

3-size-2-min-size: keep 4 in during thrashing

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

3-size-2-min-size: keep 4 in during thrashing

Workaround for 12231.

Fixes: #12231
Signed-off-by: Samuel Just <sjust@redhat.com>

Removed rhel* configurations with a goal that they will be covered in the Octo lab
and won't be used by vps runs in Sepia lab

Signed-off-by: Yuri Weinstein <yweinste@redhat.com>
(cherry picked from commit adfa1ee393c0ea0a36c0dfb8d7fda1a82f4b29ea)

suites/rbd: add EC pool variant for QEMU tests

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

Merge pull request #476 from t-miyamae/wip-10886

erasure code: added shec's initial ceph-qa-suite

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

Merge pull request #481 from dmick/master

calamari_nosetests: verify that calamari_branch exists

Merge pull request #483 from ceph/wip-excl

fix hammer-x upgrades with excl stuff in ceph_test_rados

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

Merge pull request #484 from ceph/wip_remove_rhel

Removed rhel* configurations with a goal that they will be covered in…

Removed rhel* configurations with a goal that they will be covered in the Octo lab
and won't be used by vps runs in Sepia lab

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

upgrade/hammer-x: drop symlinks to ec jobs and disable *_excl

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

suites/upgrade/hammer-x: do not pass _excl weights to ceph_test_rados

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

tasks/rados: make {write,append}_excl conditional

For upgrade tests we don't want to pass these args to (old)
ceph_test_rados.

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

calamari_nosetests: verify that calamari_branch exists

Fixes: #12206
Signed-off-by: Dan Mick <dan.mick@redhat.com>

Merge pull request #479 from ceph/wip-rgw-task

rgw: do not enable both tcp and uds for fastcgi

rgw: do not enable both tcp and uds for fastcgi

This fixes a bug where we were setting up mod_fastcgi to use both tcp
and uds.

Merge pull request #478 from ceph/wip-12172-backport-next

cephfs/test_auto_repair.py: flush journal after umount

Merge pull request #477 from ceph/wip-12172-backport-next

cephfs/test_auto_repair.py: flush journal after umount

cephfs/test_auto_repair.py: flush journal after umount

When client capabilities get released, MDS may update corresponding
inodes' client writable range and mark those inodes dirty. The auto
repair test expects MDS to trim inodes from its cache, but MDS can't
trim dirty inodes. So we should flush journal after umount.

Fixes: #12172
Signed-off-by: Yan, Zheng <zyan@redhat.com>
(cherry picked from commit 835ad8b8f3038290846c612b66f3bffcc5755338)

erasure code: added shec's initial ceph-qa-suite

Fixes: #10886(http://tracker.ceph.com/issues/10886)
Signed-off-by: Shotaro Kawaguchi <kawaguchi.s@jp.fujitsu.com>