Merge pull request #525 from ceph/wip-43448

testnode: Move dbench to epel_packages so it gets installed

Merge pull request #524 from ceph/wip-el8

testnode: el8: install gdb for test-crash.sh

testnode: el8: install gdb for test-crash.sh

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

Merge pull request #523 from ceph/wip-el7

testnode: Fix typo so task only runs on EL8

testnode: Fix typo so task only runs on EL8

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #522 from ceph/wip-bz1782899

testnode: Workaround for BZ 1782899

testnode: Workaround for BZ 1782899

This will stop all the RHEL8 failures.

Signed-off-by: David Galloway <dgallowa@redhat.com>

testnode: Move dbench to epel_packages so it gets installed

Octo lab doesn't use epel by default but by moving this package to `epel_packages`, it will get installed during this task: https://github.com/ceph/ceph-cm-ansible/blob/master/roles/testnode/tasks/yum/packages.yml#L58

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #521 from ceph/wip-exclusive

ansible_managed: Remove unmanaged keys

ansible_managed: Remove unmanaged keys

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #520 from ceph/wip-interpreter

cobbler: Don't set ansible_python_interpreter manually anymore

cobbler: Don't set ansible_python_interpreter manually anymore

Both cobbler hosts (cobbler.front.sepia.ceph.com and magna001.ceph.redhat.com) are now running versions of ansible that handle python interpreter discovery automatically.

Setting this manually was actually causing issues because some ceph-cm-ansible tasks get delegated to localhost (Like 'Clone the keys repo' in the Users role).  So when setting -e ansible_python_interpreter=/usr/bin/python3, ansible looks for that binary on the localhost.  magna001 doesn't have python3 so the playbook run would fail.

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #519 from ceph/wip-el8-nfs

testnode: el8: nfs-utils

testnode: el8: nfs-utils

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

Merge pull request #517 from ceph/wip-rhel8-kvm

testnode: el8: qemu-kvm-core + qemu-kvm-block-rbd

Merge pull request #518 from tchaikov/python3-nose-for-bionic

testnode/var/ubuntu_18: python3-nose

testnode/var/ubuntu_18: python3-nose

Signed-off-by: Kefu Chai <tchaikov@gmail.com>

testnode: el8: qemu-kvm-core + qemu-kvm-block-rbd

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

Merge pull request #516 from ceph/wip-el8

testnode: el7: install podman-docker

testnode: el7: install podman-docker

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

Merge pull request #515 from ceph/wip-el8

testnode: el8: podman-docker

testnode: el8: podman-docker

teuthology's cephadmunit.py needs to send a kill signal to the process,
and doesn't (easily) know whether to use the docker or podman command.
This let's us just use 'docker kill ...' and have it work for both.

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

Merge pull request #514 from ceph/wip-el8

testnode: el8: fix typo

testnode: el8: fix typo

sigh

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

Merge pull request #513 from ceph/wip-el8

testnode: el8: install valgrind

testnode: el8: install valgrind

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

Merge pull request #512 from ceph/wip-el8

testnode: el8: dbench, autoconf

testnode: el8: dbench, autoconf

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

Merge pull request #511 from ceph/wip-el8

testnode: el8: libacl-devel

testnode: el8: libacl-devel

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

Merge pull request #510 from ceph/wip-el8

testnode: el8: drop python3-pytz dep

testnode: el8: drop python3-pytz dep

python3-tempora is now fixed to require this.

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

Merge pull request #509 from ceph/wip-el8

testnode: el8: qemu-image

testnode: el8: qemu-img

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

Merge pull request #508 from ceph/wip-el8

testnode: el8: genisoimage

testnode: el8: genisoimage

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

Merge pull request #507 from ceph/wip-el8

testnode: python3-nose

testnode: el8: python3-virtualenv

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

testnode: python3-nose

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

Merge pull request #506 from ceph/wip-el8

temporarily require python3-pytz

temporarily require python3-pytz

This is until the python3-tempora package is fixed

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

Merge pull request #505 from ceph/wip-el8

temporarily require python3-pytz

temporarily require python3-pytz

This is until the python3-tempora package is fixed

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

Merge pull request #504 from ceph/wip-el8

el8: install gcc

el8: install gcc and git

workunits task needs these

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

Merge pull request #503 from ceph/wip-43231

testnode: Install python3 on CentOS/RHEL 7

testnode: Install python3 on CentOS/RHEL 7

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #502 from ceph/wip-everything

common: Add missing 'Everything' subdir to EPEL8 URLs

common: Add missing 'Everything' subdir to EPEL8 URLs

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #501 from ceph/wip-el8-copr

testnode: Support adding copr repos

testnode: Support adding copr repos

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #500 from ceph/wip-centos-8

testnode: lab-extras for centos_8 too

testnode: lab-extras for centos_8 too

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

Merge pull request #499 from ceph/wip-el8-extras

testnode: Add lab-extras repo to RHEL8

testnode: Add lab-extras repo to RHEL8

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #498 from kshtsk/wip-teuthology-py3

teuthology: add python3 deps for leap 15.1

teuthology: add python3 deps for leap 15.1

Signed-off-by: Kyr Shatskyy <kyrylo.shatskyy@suse.com>

Merge pull request #497 from ceph/wip-centos77

Add CentOS 7.7

Add CentOS 7.7

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #496 from ceph/wip-centos-8

testnode: install dnf-utils on el8

testnode: isntall dnf-utils on el8

Needed for package-cleanup, which is used by the kernel task.

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

Merge pull request #495 from ceph/wip-suse-kopts

cobbler: Set kernel_options for OpenSUSE distros

cobbler: Set kernel_options for OpenSUSE distros

This prevents the installer from pausing to ask for the install media

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #494 from kshtsk/wip-opensuse-rsyslog-etc

testnode: add rsyslog etc for openSUSE 15.1

testnode: add rsyslog etc for openSUSE 15.1

Signed-off-by: Kyr Shatskyy <kyrylo.shatskyy@suse.com>

Merge pull request #493 from ceph/wip-leap151

cobbler: Add OpenSUSE Leap 15.1

Merge pull request #491 from ceph/wip-fog-drive

generate-fog-csv: Support setting alternate 'Host Primary Disk' in FOG

Merge pull request #492 from kshtsk/wip-opensuse-15.1-testnodes

testnodes: add opensuse 15.1 support

cobbler: Add OpenSUSE Leap 15.1

Signed-off-by: David Galloway <dgallowa@redhat.com>

testnodes: add opensuse 15.1 support

Signed-off-by: Kyr Shatskyy <kyrylo.shatskyy@suse.com>

generate-fog-csv: Support setting alternate 'Host Primary Disk' in FOG

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #490 from ceph/wip-8-gitall

testnode: Install git on RHEL8

testnode: Install git on RHEL8

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #489 from ceph/rmcache

common: rm yum cache that doesn't get cleaned yum `yum clean`

common: rm yum cache that doesn't get cleaned with `yum clean`

Why is this even necessary?

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #488 from ceph/wip-cc-mirrors

common: Add a couple closer mirrors

common: Add a couple closer mirrors

These mirrors are in the Community Cage (where Sepia is) so they should speed yum transactions up a bit.

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #487 from ceph/wip-yumconf

common: Update yum metadata from upstream epel and exclude mirrors

common: Update yum metadata from upstream epel and exclude mirrors

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #485 from ceph/wip-28-deprecations

Refactor to resolve Deprecation Warnings

Merge pull request #486 from ceph/wip-delegate

users: delegate task to localhost

users: delegate task to localhost

TIL `connection: local` != `delegate_to: localhost`

See https://github.com/ansible/ansible/issues/16724#issuecomment-259989010

Signed-off-by: David Galloway <dgallowa@redhat.com>

cephlab: Get rid of manual python version check

Newer versions of ansible are able to connect to a machine and figure out what to set `ansible_python_interpreter` to without a stilly hack like this.

Signed-off-by: David Galloway <dgallowa@redhat.com>

users: Don't evaluate bare variables as true/false

Signed-off-by: David Galloway <dgallowa@redhat.com>

Stop using loops for package-related tasks

This is slower and will be deprecated in a future version of Ansible

Signed-off-by: David Galloway <dgallowa@redhat.com>

all: Using tests as filters is deprecated

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #484 from ceph/wip-rhel81

cobbler: Add RHEL8.1 to lab

cobbler: Add RHEL8.1 to lab

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #483 from ceph/wip-rhel78

cobbler: Add RHEL7.8

cobbler: Add RHEL7.8

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #482 from ceph/wip-skip-users

users: Tasks to allow skipping of role if all users/keys are already up to date

users: Tasks to allow skipping of role if all users/keys are already good

Signed-off-by: David Galloway <dgallowa@redhat.com>

Merge pull request #481 from ceph/Rook-Recovery-Role

Adding rook-recovery role which can be used for deploying the rook je…

Adding rook-recovery role which can be used for deploying the rook jenkins instance in the rook CI aws vpc, The deployied instance will have two ebs volumes were the jobs and the jobs configuration data is located

Signed-off-by: Adam Kraitman <akraitma@redhat.com>

Merge pull request #480 from ceph/wip-remove-multipath

testnode: remove multipath

Merge pull request #479 from ceph/wip-podman-docker

roles/testnode/vars: install podman|docker newer distros

remove opensuse leap 42.1 (very old!)

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

roles/testnode/vars: install podman|docker newer distros

This is for ceph-daemon.

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

testnode: remove multipath

multipath can interfere with krbd tests, particularly with
krbd_udev_enumerate.yaml.  Much like udev, multipath opens block
devices behind the scenes, making the block device busy when it
shouldn't be.

None of the lab tests depend on multipath (it's not installed on
ubuntu), so just remove it rather than messing with /etc/multipath.conf
and restarting multipathd.  If the need for multipath comes up, rbd
devices would need to be blacklisted like so:

  blacklist {
          devnode "^rbd[0-9]+"
  }

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

Merge pull request #478 from ceph/wip-ifdown

cobbler: Allow ifdown to fail during rc.local