]>
git.apps.os.sepia.ceph.com Git - ceph-cm-ansible.git/log
Ken Dreyer [Thu, 12 Nov 2015 19:18:12 +0000 (12:18 -0700)]
testnode: do not install fastcgi on RHEL
Red Hat QE occasionally tests full in-place OS upgrades from RHEL 6 to
RHEL 7. When mod_fastcgi is pre-installed on a RHEL 6 node, and the node
is upgraded to RHEL 7, then Apache will refuse to start unless the QE
team member also upgrades the mod_fastcgi package explicitly. (see
https://bugzilla.redhat.com/
1230679 for an example)
Since RGW's Civetweb is becoming mainstream, we don't really need to
focus on mod_fastcgi for Ceph any more, and having it pre-installed on
our RHEL nodes is more of a hindrance than a help.
I'm leaving the CentOS configs alone, for now, so we don't interfere
with any upstream testing with mod_fastcgi.
Andrew Schoen [Fri, 6 Nov 2015 17:29:08 +0000 (11:29 -0600)]
Merge pull request #165 from ceph/wip-paddles-nginx
Use nginx as a reverse proxy for paddles
Zack Cerza [Fri, 6 Nov 2015 17:19:43 +0000 (10:19 -0700)]
Use nginx as a reverse proxy for paddles
This is to avoid timeouts when gunicorn workers are busy.
Signed-off-by: Zack Cerza <zack@redhat.com>
Andrew Schoen [Wed, 4 Nov 2015 17:49:04 +0000 (11:49 -0600)]
Merge pull request #153 from dillaman/wip-13308
testnode: install QEMU packages on Fedora, RHEL, and CentOS
Zack Cerza [Tue, 3 Nov 2015 21:28:36 +0000 (14:28 -0700)]
Merge pull request #160 from dmick/master
vmlist.py: query multiple openstack providers, 'regions' in provider, add .conf template
Dan Mick [Tue, 3 Nov 2015 21:26:52 +0000 (13:26 -0800)]
Merge pull request #163 from ceph/wip-correct-hostname
testnodes: Set hostnames properly
Reviewed-by: Dan Mick <dmick@redhat.com>
Zack Cerza [Tue, 3 Nov 2015 21:15:17 +0000 (14:15 -0700)]
testnodes: Set hostnames properly
Instead of using convoluted logic to determine whether or not to set the
hostname, just do it. The hostname module is nice and idempotent.
This was inspired by http://tracker.ceph.com/issues/13679
Signed-off-by: Zack Cerza <zack@redhat.com>
Dan Mick [Mon, 2 Nov 2015 23:57:13 +0000 (15:57 -0800)]
vmlist.py: remove some broken hosts from the default list
Signed-off-by: Dan Mick <dan.mick@redhat.com>
Dan Mick [Tue, 27 Oct 2015 06:34:08 +0000 (23:34 -0700)]
vmlist.py: Add prototype .vmlist.conf file
Signed-off-by: Dan Mick <dan.mick@redhat.com>
Dan Mick [Tue, 27 Oct 2015 05:57:36 +0000 (22:57 -0700)]
Add cloud-<provider-string> sections, and list them all
For dreamcompute, project id is enough; for ovh, tenant id
and region id need to be specified. Allow those optional
fields, and loop across all providers; for regions, take a list
and subloop across all regions (if it exists).
For ovh, this means adding a config section like
[cloud-ovh-cattle]
cloud_user: <user>
cloud_password: <pass>
cloud_project_id:
5633955729735406
cloud_tenant_id:
131b886b156a4f84b5f41baf2fbe646c
cloud_region_names: GRA1, BHS1
cloud_auth_url: https://auth.cloud.ovh.net/v2.0
Signed-off-by: Dan Mick <dan.mick@redhat.com>
Dan Mick [Tue, 27 Oct 2015 05:51:57 +0000 (22:51 -0700)]
vmlist.py: change from 'defaults' to a 'global' section
defaults adds the entries to every section, which is wasteful. Not
thrilled with managing my own defaults, but it seems easy enough
to make [global] section and prepopulate it (in preparation for more
sections). Also allow 'get' to return None for nonexistent key.
Signed-off-by: Dan Mick <dan.mick@redhat.com>
Zack Cerza [Thu, 29 Oct 2015 22:43:12 +0000 (16:43 -0600)]
Merge pull request #162 from ceph/wip-dban-autonuke
cobbler: renaming cobbler distro/profile to indicate autonuke/non-interactive
dgalloway [Thu, 29 Oct 2015 21:50:12 +0000 (17:50 -0400)]
cobbler: renaming cobbler distro/profile to indicate autonuke/non-interactive
Signed-off-by: dgalloway <dgallowa@REDHAT.COM>
Zack Cerza [Thu, 29 Oct 2015 19:25:10 +0000 (13:25 -0600)]
Merge pull request #143 from ceph/wip-cobbler-hostname
Set hostname in cobbler system profiles
Dan Mick [Tue, 27 Oct 2015 21:00:42 +0000 (14:00 -0700)]
Merge pull request #159 from ceph/wip-teuthology-max-job-time
Allow setting max job time for teuthology
Reviewed-by: Dan Mick <dmick@redhat.com>
Zack Cerza [Tue, 27 Oct 2015 20:34:20 +0000 (14:34 -0600)]
Allow setting max job time for teuthology
Signed-off-by: Zack Cerza <zack@redhat.com>
Andrew Schoen [Tue, 27 Oct 2015 17:06:38 +0000 (12:06 -0500)]
Merge pull request #158 from ceph/wip-testnodes-disk-group
testnodes: Add teuthology_user to 'disk' group
Zack Cerza [Mon, 26 Oct 2015 22:39:15 +0000 (16:39 -0600)]
testnodes: Add teuthology_user to 'disk' group
We were already doing this for apt systems; do it for yum systems as
well
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Fri, 23 Oct 2015 15:44:30 +0000 (09:44 -0600)]
Merge pull request #157 from dachary/wip-blkin
Temporarily disable blkin repo (part 2)
Loic Dachary [Fri, 23 Oct 2015 09:43:57 +0000 (11:43 +0200)]
Temporarily disable blkin repo (part 2)
Tests that need blkin are welcome to install it in a way that makes sense.
http://tracker.ceph.com/issues/13572
Signed-off-by: Loic Dachary <ldachary@redhat.com>
Andrew Schoen [Thu, 22 Oct 2015 20:17:35 +0000 (15:17 -0500)]
Merge pull request #156 from ceph/wip-drop-blkin
Temporarily disable blkin repo
Zack Cerza [Thu, 22 Oct 2015 20:13:41 +0000 (14:13 -0600)]
Temporarily disable blkin repo
http://tracker.ceph.com/issues/13572
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Thu, 15 Oct 2015 19:52:19 +0000 (13:52 -0600)]
Merge pull request #155 from ceph/wip-puddle-updates
puddle: nginx configs for HTTPS
Ken Dreyer [Wed, 14 Oct 2015 22:59:43 +0000 (16:59 -0600)]
puddle: nginx Basic auth for /ubuntu
The production web server that hosts Ubuntu repositories will be secured
with HTTP Basic auth. Simulate this on the puddle web server as well.
/etc/nginx/htpasswd is managed outside of Ansible.
Ken Dreyer [Wed, 14 Oct 2015 20:37:01 +0000 (14:37 -0600)]
puddle: nginx configs for HTTPS
I'm adding HTTPS to our Puddle web server in order to support test Apt
repos over HTTPS.
The packages will be hosted under http://puddle.fqdn/ubuntu/
The key and certificate files are managed outside of Ansible.
Andrew Schoen [Tue, 13 Oct 2015 13:24:25 +0000 (08:24 -0500)]
Merge pull request #154 from ceph/wip-paddles
paddles: Update prod.py
Zack Cerza [Mon, 12 Oct 2015 23:12:32 +0000 (17:12 -0600)]
paddles: Update prod.py
Fixes a bug introduced by:
https://github.com/pecan/pecan/commit/
b14f4aa1682b77681a760ed948f8972a4e0d7d2a
Signed-off-by: Zack Cerza <zack@redhat.com>
Jason Dillaman [Mon, 12 Oct 2015 20:25:11 +0000 (16:25 -0400)]
testnode: install QEMU packages on Fedora, RHEL, and CentOS
Several RBD test cases require QEMU to be present to verify integration.
Ubuntu package list already included QEMU packages.
Fixes: #13308
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
Dan Mick [Fri, 9 Oct 2015 21:52:39 +0000 (14:52 -0700)]
Merge pull request #152 from ceph/wip-paddles
paddles: Always include secrets
Reviewed-by: Dan Mick <dmick@redhat.com>
Zack Cerza [Fri, 9 Oct 2015 21:49:01 +0000 (15:49 -0600)]
paddles: Always include secrets
Currently this only stores the db_pass var. Also fix the name of the
play.
Signed-off-by: Zack Cerza <zack@redhat.com>
Dan Mick [Fri, 9 Oct 2015 21:32:46 +0000 (14:32 -0700)]
Merge pull request #151 from ceph/wip-ovh
Various changes for OVH deployment
Reviewed-by: Dan Mick <dmick@redhat.com>
Zack Cerza [Fri, 9 Oct 2015 21:13:44 +0000 (15:13 -0600)]
teuthology: Add teuthology_yaml_extra
This is an optional var that allows arbitrary yaml fragments to be
included in teuthology.yaml
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Fri, 9 Oct 2015 21:13:24 +0000 (15:13 -0600)]
teuthology: Remove bogus default
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Fri, 9 Oct 2015 21:12:51 +0000 (15:12 -0600)]
teuthology: Serve logs under /teuthology
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Fri, 9 Oct 2015 21:10:57 +0000 (15:10 -0600)]
teuthology: Add tags for config and logs
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Fri, 9 Oct 2015 21:10:21 +0000 (15:10 -0600)]
paddles: Restart if prod.py changes
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Fri, 9 Oct 2015 21:09:42 +0000 (15:09 -0600)]
paddles: Always call the set_fact plays
Signed-off-by: Zack Cerza <zack@redhat.com>
Dan Mick [Fri, 9 Oct 2015 17:43:56 +0000 (10:43 -0700)]
Merge pull request #150 from ceph/wip-fix-users
Don't attempt to add groups that don't exist
Reviewed-by: Dan Mick <dmick@redhat.com>
Zack Cerza [Fri, 9 Oct 2015 16:47:13 +0000 (10:47 -0600)]
Don't attempt to add groups that don't exist
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Fri, 9 Oct 2015 16:08:16 +0000 (10:08 -0600)]
Merge pull request #149 from dmick/master
Create users with same-named primary group
Dan Mick [Fri, 9 Oct 2015 00:41:17 +0000 (17:41 -0700)]
Create users with same-named primary group
If adding sudo, add as secondary group ('groups:'). This is more
standard, and allows for quick enumeration of "who's in sudo" by
examining /etc/group
Signed-off-by: Dan Mick <dan.mick@redhat.com>
Andrew Schoen [Tue, 6 Oct 2015 16:23:35 +0000 (11:23 -0500)]
Merge pull request #139 from ceph/wip-dban
Add dummy dban-2.3.0 entry
Andrew Schoen [Tue, 6 Oct 2015 16:23:06 +0000 (11:23 -0500)]
Merge pull request #148 from ceph/wip-teuthology
teuthology: Serve logs over http
Zack Cerza [Fri, 2 Oct 2015 22:35:20 +0000 (16:35 -0600)]
teuthology: Serve logs over http
Signed-off-by: Zack Cerza <zack@redhat.com>
Andrew Schoen [Fri, 2 Oct 2015 16:50:44 +0000 (11:50 -0500)]
Merge pull request #147 from ceph/wip-teuthology
teuthology: manage teuthology system users
Zack Cerza [Tue, 29 Sep 2015 19:45:03 +0000 (13:45 -0600)]
teuthology: manage teuthology system users
'teuthology' for scheduling and 'teuthworker' for execution
This is intended to mirror our existing setups
Signed-off-by: Zack Cerza <zack@redhat.com>
Andrew Schoen [Fri, 2 Oct 2015 14:13:50 +0000 (09:13 -0500)]
Merge pull request #138 from ceph/wip-puddle-updates
puddle: multiple updates
Andrew Schoen [Thu, 1 Oct 2015 22:46:24 +0000 (17:46 -0500)]
Merge pull request #146 from ceph/wip-paddles
Add new paddles role
Zack Cerza [Thu, 1 Oct 2015 16:53:40 +0000 (10:53 -0600)]
pulpito: Restart if the config changes
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Wed, 30 Sep 2015 22:38:49 +0000 (16:38 -0600)]
paddles: Add new paddles role
Signed-off-by: Zack Cerza <zack@redhat.com>
dgalloway [Fri, 25 Sep 2015 01:32:45 +0000 (21:32 -0400)]
Set hostname in cobbler system profiles
This should take care of a problem we encountered in issue 12672
http://tracker.ceph.com/issues/12672
Signed-off-by: dgalloway <dgallowa@REDHAT.COM>
Zack Cerza [Thu, 1 Oct 2015 16:36:58 +0000 (10:36 -0600)]
Merge pull request #145 from ceph/wip-hostname
testnode: set_hostname should be a global task
dgalloway [Wed, 30 Sep 2015 20:27:58 +0000 (16:27 -0400)]
Update documentation for lab_domain var
dgalloway [Wed, 30 Sep 2015 19:11:26 +0000 (15:11 -0400)]
testnode: set_hostname should be a global task
Signed-off-by: dgalloway <dgallowa@REDHAT.COM>
Ken Dreyer [Mon, 21 Sep 2015 18:17:05 +0000 (12:17 -0600)]
puddle: new RH-COMMON advisory undergoing testing
Adjust the RHEL 6 and 7 RH-COMMON Puddle configs for a new async update
that will be undergoing testing.
Ken Dreyer [Fri, 18 Sep 2015 22:43:47 +0000 (16:43 -0600)]
puddle: output RHCEPH 1.3 to separate locations
I periodically generate Puddles for both RHCS 1.3.0-async and 1.3.1.
It's confusing when these Puddles exist in the same parent directory.
Split them up into two directories so it is easier to understand what
each Puddle is.
Andrew Schoen [Fri, 25 Sep 2015 19:54:11 +0000 (14:54 -0500)]
Merge pull request #144 from ceph/wip-13237
Remove custom qemu packages
dgalloway [Fri, 25 Sep 2015 19:46:41 +0000 (15:46 -0400)]
Remove custom qemu packages
Signed-off-by: dgalloway <dgallowa@REDHAT.COM>
Dan Mick [Thu, 24 Sep 2015 22:33:46 +0000 (15:33 -0700)]
Merge pull request #132 from ceph/wip-magna-smart
Updates to nagios disk monitoring script
Reviewed-by: Dan Mick <dmick@redhat.com>
Zack Cerza [Thu, 24 Sep 2015 21:12:45 +0000 (15:12 -0600)]
Merge pull request #142 from ceph/wip-12739
Use local epel mirrorlist
dgalloway [Thu, 24 Sep 2015 16:16:15 +0000 (12:16 -0400)]
Configure yum to use local epel mirrorlists
Signed-off-by: dgalloway <dgallowa@REDHAT.COM>
Zack Cerza [Wed, 23 Sep 2015 21:59:39 +0000 (15:59 -0600)]
Merge pull request #141 from ceph/wip-rm-cuttlefish-repo
testnode: rm cuttlefish repository
Ken Dreyer [Wed, 23 Sep 2015 21:52:14 +0000 (15:52 -0600)]
testnode: rm cuttlefish repository
Prior to this commit, every RHEL 6 and CentOS 6 node would automatically
have the cuttlefish repository enabled.
This hardcoding seems wrong. Now that we've moved the RPMs off of
ceph.com to download.ceph.com, it actually breaks nodes. So let's stop
hard-coding this one altogether.
Reported-by: Warren Usui <wusui@redhat.com>
Zack Cerza [Wed, 23 Sep 2015 20:16:51 +0000 (14:16 -0600)]
Merge pull request #140 from ceph/wip-13203-epel-fcgi
testnode: add fcgi to the list of EPEL packages
Ken Dreyer [Wed, 23 Sep 2015 19:24:17 +0000 (13:24 -0600)]
testnode: add fcgi to the list of EPEL packages
For Ceph upstream, we rely on fcgi from EPEL.
http://tracker.ceph.com/issues/13203 Refs: #13203
Zack Cerza [Tue, 22 Sep 2015 19:25:25 +0000 (13:25 -0600)]
Add dummy dban-2.3.0 entry
Signed-off-by: Zack Cerza <zack@redhat.com>
Andrew Schoen [Mon, 21 Sep 2015 20:23:49 +0000 (15:23 -0500)]
Merge pull request #135 from ceph/wip-cobbler-rescue
cobbler_profile: add non-iso image support and inktank-rescue
dgalloway [Mon, 21 Sep 2015 17:23:09 +0000 (13:23 -0400)]
Make multiline output default
dgalloway [Mon, 21 Sep 2015 17:25:02 +0000 (13:25 -0400)]
Found typo in conditional
Zack Cerza [Mon, 21 Sep 2015 16:22:18 +0000 (10:22 -0600)]
Merge pull request #137 from ceph/rm-ceph-extras
rm ceph-extras repository config
Zack Cerza [Mon, 21 Sep 2015 16:18:42 +0000 (10:18 -0600)]
Merge pull request #136 from dmick/master
Some fixes to vmlist.py: update host lists, fix error from missing cloud config items
Ken Dreyer [Mon, 21 Sep 2015 15:52:21 +0000 (09:52 -0600)]
rm ceph-extras repository config
This repository contains packages that we don't have time to continue
maintaining. In particular, the CentOS 6 QEMU packages were vulnerable
to VENOM (CVE-2015-3456) and no users should rely on them. The other
packages in "ceph-extras" will continue to incur security issues over
time.
With the recent migration from "ceph.com" -> "download.ceph.com", we've
chosen not to copy these "ceph-extras" files over to the new locations.
Dan Mick [Sat, 19 Sep 2015 06:03:21 +0000 (23:03 -0700)]
vmlist.py: default cloud_* so their absence doesn't cause errors
Signed-off-by: Dan Mick <dan.mick@redhat.com>
Dan Mick [Sat, 19 Sep 2015 06:02:57 +0000 (23:02 -0700)]
vmlist.py: senta01 was inexplicably missing (and made me miss ursula!)
Signed-off-by: Dan Mick <dan.mick@redhat.com>
Dan Mick [Sat, 19 Sep 2015 06:02:08 +0000 (23:02 -0700)]
vmlist.py: mira004 is back from the dead
Signed-off-by: Dan Mick <dan.mick@redhat.com>
Zack Cerza [Thu, 17 Sep 2015 00:31:25 +0000 (18:31 -0600)]
Add distro_prep tag to distro_prep.yml
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Wed, 16 Sep 2015 23:30:04 +0000 (17:30 -0600)]
Add dummy inktank-rescue entry
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Wed, 16 Sep 2015 23:01:20 +0000 (17:01 -0600)]
Add inktank-rescue to cobbler.yml
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Tue, 15 Sep 2015 21:23:02 +0000 (15:23 -0600)]
cobbler_profile: Import kernel/initrd distros
This will be useful for things like rescue images.
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Tue, 15 Sep 2015 21:21:45 +0000 (15:21 -0600)]
cobbler_profile: Split out ISO importing
So that we can have a separate codepath for importing distros from
non-ISO sources.
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Tue, 15 Sep 2015 21:18:51 +0000 (15:18 -0600)]
Cobbler: Create other_image_dir
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Tue, 15 Sep 2015 20:49:23 +0000 (14:49 -0600)]
Cobbler: Add other_image_dir default
For storing non-iso images for distro/profile creation
Signed-off-by: Zack Cerza <zack@redhat.com>
Andrew Schoen [Wed, 16 Sep 2015 17:52:09 +0000 (12:52 -0500)]
Merge pull request #134 from ceph/wip-fix-distros
cobbler_profile fixes
Zack Cerza [Wed, 16 Sep 2015 17:45:52 +0000 (11:45 -0600)]
Ubuntu 14.04.2 -> 14.04.3
14.04.2 is deprecated and ISOs are no longer shipped.
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Wed, 16 Sep 2015 17:30:33 +0000 (11:30 -0600)]
CentOS 6.6 -> 6.7
6.6 is deprecated and ISOs are no longer shipped.
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Wed, 16 Sep 2015 17:19:18 +0000 (11:19 -0600)]
Add dummy entry for RHEL-6.7
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Wed, 16 Sep 2015 17:19:02 +0000 (11:19 -0600)]
Don't attempt to manage tftpd on apt systems
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Wed, 16 Sep 2015 17:18:41 +0000 (11:18 -0600)]
Revert "Simplify extra_distros logic"
This reverts commit
0e5eaacba746c698315755b2f8464043d8f37910 .
Andrew Schoen [Wed, 16 Sep 2015 15:35:05 +0000 (10:35 -0500)]
Merge pull request #133 from ceph/wip-pulpito-playbook
Adds the pulpito playbook
Andrew Schoen [Wed, 16 Sep 2015 15:33:55 +0000 (10:33 -0500)]
Adds the pulpito playbook
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
dgalloway [Wed, 16 Sep 2015 14:51:17 +0000 (10:51 -0400)]
Some useful comments explaining functions
dgalloway [Wed, 16 Sep 2015 14:28:04 +0000 (10:28 -0400)]
Count num failed drives and output in first line of multiline output
dgalloway [Wed, 16 Sep 2015 14:16:39 +0000 (10:16 -0400)]
Proceed as JBOD if not RAID
Andrew Schoen [Wed, 16 Sep 2015 14:09:16 +0000 (09:09 -0500)]
Merge pull request #131 from ceph/wip-pulpito
Pulpito role
Zack Cerza [Wed, 9 Sep 2015 18:46:04 +0000 (12:46 -0600)]
Add pulpito role to cephlab.yml
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Wed, 9 Sep 2015 18:45:42 +0000 (12:45 -0600)]
Add role to deploy pulpito
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Wed, 9 Sep 2015 16:38:25 +0000 (10:38 -0600)]
Merge pull request #130 from ceph/wip-12921
downstream-setup: yum_repos will now be removed during cleanup
Andrew Schoen [Wed, 9 Sep 2015 15:07:45 +0000 (10:07 -0500)]
downstream-setup: yum_repos will now be removed during cleanup
Fixes: http://tracker.ceph.com/issues/12921
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
djgalloway [Tue, 8 Sep 2015 23:49:13 +0000 (16:49 -0700)]
Merge pull request #129 from ceph/wip-smart
smart.pl multiline mode: show failed/missing drives as well
Reviewed-by: David Galloway <dgalloway@redhat.com>
Dan Mick [Tue, 8 Sep 2015 23:41:18 +0000 (16:41 -0700)]
smart.pl: drive numbers are 1-based
Signed-off-by: Dan Mick <dan.mick@redhat.com>