]>
git.apps.os.sepia.ceph.com Git - teuthology.git/log
Loic Dachary [Mon, 14 Sep 2015 12:20:57 +0000 (14:20 +0200)]
openstack: clone from the teuthology master
Do not clone from an obsolete development branch by default.
Signed-off-by: Loic Dachary <loic@dachary.org>
Zack Cerza [Tue, 8 Sep 2015 19:16:03 +0000 (13:16 -0600)]
Merge pull request #618 from dachary/wip-openstack-throttle
openstack: throttle job scheduling by default
Zack Cerza [Tue, 8 Sep 2015 19:14:55 +0000 (13:14 -0600)]
Merge pull request #619 from dachary/wip-openstack-jessie
openstack: add jessie / debian 8 support
Loic Dachary [Mon, 7 Sep 2015 21:32:56 +0000 (23:32 +0200)]
openstack: add jessie / debian 8 support
http://tracker.ceph.com/issues/12988 Fixes: #12988
Signed-off-by: Loic Dachary <loic@dachary.org>
Loic Dachary [Mon, 7 Sep 2015 11:42:19 +0000 (13:42 +0200)]
openstack: throttle job scheduling by default
When running a suite with 50 jobs, it will schedule about 100 instance
creation within less than a minute. It is likely to exceed the API
quotas of the OpenStack provider (number of instance creation per
minute) and lead to instance creation failures.
Set the teuthology-suite to be called with --throttle 15 by default,
that is about 60 / 15 = 4 * 2 = ~8 server creation per minute.
http://tracker.ceph.com/issues/12977 Fixes: #12977
Signed-off-by: Loic Dachary <loic@dachary.org>
Loic Dachary [Mon, 7 Sep 2015 11:40:47 +0000 (13:40 +0200)]
suite: add --throttle to pause between jobs
When scheduling, wait SLEEP seconds between jobs. Useful to avoid
bursts that may be too hard on the underlying infrastructure or exceed
OpenStack API limits (server creation per minute for instance).
http://tracker.ceph.com/issues/12977 Refs: #12977
Signed-off-by: Loic Dachary <loic@dachary.org>
vasukulkarni [Tue, 8 Sep 2015 17:40:11 +0000 (10:40 -0700)]
Merge pull request #620 from ceph/wip-12995
Fix bug in using priority 0
Zack Cerza [Tue, 8 Sep 2015 17:32:54 +0000 (11:32 -0600)]
Fix bug in using priority 0
http://tracker.ceph.com/issues/12995
Signed-off-by: Zack Cerza <zack@redhat.com>
Loic Dachary [Mon, 7 Sep 2015 11:38:12 +0000 (13:38 +0200)]
openstack: fix teuthology config test for clone
The test failed because it looked for upload= instead of clone=
Signed-off-by: Loic Dachary <ldachary@dachary.org>
Sage Weil [Sun, 6 Sep 2015 12:33:41 +0000 (08:33 -0400)]
Revert "use modern paramiko"
This reverts commit
76e1b093354dd31d94dd56299d1a245ff82db72e .
I see lots of these failures:
2015-09-06T00:02:43.199 ERROR:paramiko.transport:SSHException: Key-exchange timed out waiting for key negotiation
Dan Mick [Fri, 4 Sep 2015 23:03:07 +0000 (16:03 -0700)]
Merge pull request #610 from ceph/wip-rh131-cephlab
Add version check for 1.3.1, add ansible task
Reviewed-by: Dan Mick <dmick@redhat.com>
Vasu Kulkarni [Thu, 3 Sep 2015 16:23:50 +0000 (12:23 -0400)]
Add ansible task to install repo if config provides it
If playbook is specificed as shown in doc string usage, It will run ansible task and install
the repos so that rpm's can be installed from the provided repo
Signed-off-by: Vasu Kulkarni <vasu@redhat.com>
Vasu Kulkarni [Thu, 3 Sep 2015 16:10:39 +0000 (12:10 -0400)]
Don't consume config['rhbuild'] until overrides are merged
rhbuild from config was read before overrides was merged
Signed-off-by: Vasu Kulkarni <vasu@redhat.com>
Zack Cerza [Fri, 4 Sep 2015 18:57:56 +0000 (12:57 -0600)]
Merge pull request #616 from ceph/wip-modern-paramiko
use modern paramiko
Sage Weil [Fri, 4 Sep 2015 18:55:39 +0000 (14:55 -0400)]
use modern paramiko
We need >= 1.15 to connect to newer OpenSSH, such as the version used
by debian jessie.
Signed-off-by: Sage Weil <sage@redhat.com>
Vasu Kulkarni [Thu, 3 Sep 2015 16:01:51 +0000 (12:01 -0400)]
Add support for rh version 1.3.1
This adds support for RH version 1.3.1, previously it will work
with only 1.3.0 downstream builds
Signed-off-by: Vasu Kulkarni <vasu@redhat.com>
Zack Cerza [Thu, 3 Sep 2015 17:46:46 +0000 (11:46 -0600)]
Merge pull request #615 from dachary/wip-openstack-doc
openstack: update HOWTO
Loic Dachary [Thu, 3 Sep 2015 16:54:18 +0000 (18:54 +0200)]
openstack: update HOWTO
Clone from master instead of a designated branch.
Signed-off-by: Loic Dachary <loic@dachary.org>
Zack Cerza [Thu, 3 Sep 2015 16:25:07 +0000 (10:25 -0600)]
Remove libpython-dev for now
It's not available on precise and is breaking testing in the sepia lab
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Thu, 3 Sep 2015 15:48:46 +0000 (09:48 -0600)]
Merge pull request #613 from ceph/wip-12934
GitbuilderProject: fetch sha1 from gitbuilder if not in the config
Andrew Schoen [Thu, 3 Sep 2015 15:12:12 +0000 (10:12 -0500)]
GitbuilderProject: fetch sha1 from gitbuilder if not in the config
If the sha1 is not provided in the job config, it needs to be fetched
from gitbuilder.
See: http://tracker.ceph.com/issues/12934
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Zack Cerza [Wed, 2 Sep 2015 22:16:49 +0000 (16:16 -0600)]
Merge pull request #592 from dachary/wip-6502-openstack-v3
transparent OpenStack provisioning for teuthology-suite
Loic Dachary [Wed, 2 Sep 2015 22:05:35 +0000 (00:05 +0200)]
openstack: explains args vs argv
Signed-off-by: Loic Dachary <loic@dachary.org>
Loic Dachary [Mon, 31 Aug 2015 21:18:15 +0000 (23:18 +0200)]
openstack: --upload defaults to teuthology-logs.public.ceph.com
Instead of a non-ceph domain.
Signed-off-by: Loic Dachary <loic@dachary.org>
Zack Cerza [Thu, 27 Aug 2015 22:03:35 +0000 (16:03 -0600)]
Add unit tests for OpenStack config defaults
Signed-off-by: Zack Cerza <zack@redhat.com>
Loic Dachary [Fri, 28 Aug 2015 14:03:02 +0000 (16:03 +0200)]
openstack: make git clone configurable
When teuthology-openstack clone theuthology for the purpose of creating
the cluster, use the clone configuration value instead of a hardcoded value.
Signed-off-by: Loic Dachary <loic@dachary.org>
Loic Dachary [Fri, 28 Aug 2015 13:38:07 +0000 (15:38 +0200)]
openstack: set config default and documentation
Add documentation for the ~/.teuthology.yaml OpenStack configuration
values and set reasonable defaults.
Signed-off-by: Loic Dachary <loic@dachary.org>
Loic Dachary [Tue, 25 Aug 2015 08:17:05 +0000 (10:17 +0200)]
openstack: remove debug line that does not print
It misses the IP that was set previous to
20bdef4b7892178d5e0fb8fa23beae52ba143eb2
Signed-off-by: Loic Dachary <loic@dachary.org>
Zack Cerza [Fri, 21 Aug 2015 16:20:25 +0000 (10:20 -0600)]
OpenStack(): Don't connect via paramiko directly
Instead, use teuthology.orchestra.connection.connect() - to take
advantage of its parsing of the ssh config. This way users won't run
into surprises when using OpenStack vs. other machine types.
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Fri, 21 Aug 2015 16:16:02 +0000 (10:16 -0600)]
connect(): Optionally override key_filename
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Fri, 21 Aug 2015 16:06:47 +0000 (10:06 -0600)]
connect(): Make retries optional
Signed-off-by: Zack Cerza <zack@redhat.com>
Loic Dachary [Thu, 20 Aug 2015 22:57:56 +0000 (00:57 +0200)]
openstack: teuthology.init needs credentials to nuke
And it cannot assume it is running from the home directory of the user
although it's the case most of the time.
Signed-off-by: Loic Dachary <loic@dachary.org>
Loic Dachary [Wed, 19 Aug 2015 00:50:06 +0000 (02:50 +0200)]
openstack: centos-7 image must be recent
Using the CentOS 7 image available at
http://cloud.centos.org/centos/7/images/CentOS-7-x86_64-GenericCloud.qcow2
ansible will think it runs
RedHat 7.1.1503 (Core)
and with a more recent image
http://cloud.centos.org/centos/7/images/CentOS-7-x86_64-GenericCloud-1503.qcow2
it will think that it runs
CentOS Linux release 7.1.1503 (Core)
http://tracker.ceph.com/issues/12725 Fixes: #12725
Signed-off-by: Loic Dachary <loic@dachary.org>
tmp
Loic Dachary [Sun, 9 Aug 2015 19:59:31 +0000 (21:59 +0200)]
upload teuthology archive on completion
Signed-off-by: Loic Dachary <loic@dachary.org>
Loic Dachary [Sun, 19 Jul 2015 06:55:48 +0000 (08:55 +0200)]
transparent OpenStack provisioning for teuthology-suite
The teuthology-openstack command is a wrapper around teuthology-suite
that transparently creates the teuthology cluster using OpenStack
virtual machine.
For machines of machine_type == openstack in paddles, when locking a
machine, an instance is created in the matching OpenStack cluster with:
openstack server create redhat
And renamed into redhat042010 if assigned the IP x.x.42.10/16. It is then
locked in paddles which has been prepare with one slot for each
available IP in the range.
An OpenStack cluster is defined in the .teuthology.yaml file as follows:
openstack:
user-data: teuthology/openstack/openstack-{os_type}-{os_version}-user-data.txt
nameserver: 167.114.252.136
machine:
disk: 10 # GB
ram: 7000 # MB
cpus: 1
volumes:
count: 0
size: 1 # GB
flavor-select-regexp: ^vps-ssd
subnet: 167.114.224.0/19
When the machine is unlocked, it is destroyed with
openstack server delete redhat042010
The python-openstackclient command line is used instead of the
corresponding API because it is well maintained and documented.
Integration tests require an OpenStack tenant.
http://tracker.ceph.com/issues/6502 Fixes: #6502
Signed-off-by: Loic Dachary <loic@dachary.org>
Alfredo Deza [Wed, 2 Sep 2015 14:06:13 +0000 (10:06 -0400)]
Merge pull request #608 from ceph/wip-jessie
packaging: add jessie codename
Sage Weil [Wed, 2 Sep 2015 13:57:37 +0000 (09:57 -0400)]
packaging: add jessie codename
Signed-off-by: Sage Weil <sage@redhat.com>
Andrew Schoen [Mon, 31 Aug 2015 21:48:49 +0000 (16:48 -0500)]
Merge pull request #607 from ceph/wip-suite-very-verbose
teuthology-suite: very-verbose mode
Zack Cerza [Mon, 31 Aug 2015 21:26:00 +0000 (15:26 -0600)]
teuthology-suite: very-verbose mode
If called with --dry-run --verbose --verbose (or --dry-run -vv), call
teuthology-schedule --dry-run for each job that would be scheduled. This
causes teuthology-schedule to print out each would-be job config.
Signed-off-by: Zack Cerza <zack@redhat.com>
Loic Dachary [Mon, 31 Aug 2015 15:30:54 +0000 (17:30 +0200)]
Merge pull request #606 from dachary/wip-11881-multipath
internal: ceph-deploy may have ceph-create-keys INFO in syslog
Andrew Schoen [Mon, 31 Aug 2015 14:46:25 +0000 (09:46 -0500)]
Merge pull request #605 from ceph/wip-ceph-deploy-commit
suite: pass dev-commit: <sha1> to ceph-deploy
Loic Dachary [Fri, 28 Aug 2015 21:29:35 +0000 (23:29 +0200)]
internal: ceph-deploy may have ceph-create-keys INFO in syslog
This is not a concern for error and should be ignored.
Signed-off-by: Loic Dachary <loic@dachary.org>
Zack Cerza [Fri, 28 Aug 2015 20:55:05 +0000 (14:55 -0600)]
Merge pull request #604 from ceph/wip-fix-kernel-centos-version
packaging: only mangle the version string for the ceph project
Sage Weil [Fri, 28 Aug 2015 20:52:06 +0000 (16:52 -0400)]
suite: pass dev-commit: <sha1> to ceph-deploy
Use the new ceph-deploy install --dev-commit <sha1> option to inform the
install.
Signed-off-by: Sage Weil <sage@redhat.com>
Andrew Schoen [Fri, 28 Aug 2015 20:07:18 +0000 (15:07 -0500)]
packaging: only mangle the version string for the ceph project
This parsing is not needed on other projects and it was breaking
the kernel task by trying to verify an incorrect version was installed.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Zack Cerza [Wed, 26 Aug 2015 22:32:04 +0000 (16:32 -0600)]
Merge pull request #603 from ceph/wip-12795
kernel: use packaging.GitbuilderProject to talk with gitbuilder.
Andrew Schoen [Wed, 26 Aug 2015 16:45:10 +0000 (11:45 -0500)]
kernel: use packaging.GitbuilderProject to talk with gitbuilder.
This fixes issue: http://tracker.ceph.com/issues/12795
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Andrew Schoen [Mon, 24 Aug 2015 19:25:20 +0000 (14:25 -0500)]
Merge pull request #602 from ceph/wip-show-job-priority
When listing jobs, show their priority
Zack Cerza [Mon, 24 Aug 2015 17:37:59 +0000 (11:37 -0600)]
When listing jobs, show their priority
Signed-off-by: Zack Cerza <zack@redhat.com>
Andrew Schoen [Fri, 21 Aug 2015 17:09:52 +0000 (12:09 -0500)]
Merge pull request #601 from ceph/wip-ansible-test
test_execute_playbook(): patch Remote.reconnect()
Zack Cerza [Fri, 21 Aug 2015 16:56:42 +0000 (10:56 -0600)]
test_execute_playbook(): patch Remote.reconnect()
Speeds up that unit test greatly.
Signed-off-by: Zack Cerza <zack@redhat.com>
Sage Weil [Thu, 20 Aug 2015 13:29:26 +0000 (09:29 -0400)]
valgrind.supp: ignore boost::uuid randomness from uninit memory
https://svn.boost.org/trac/boost/ticket/7248
Signed-off-by: Sage Weil <sage@redhat.com>
Zack Cerza [Wed, 19 Aug 2015 16:14:47 +0000 (10:14 -0600)]
Merge pull request #600 from ceph/wip-ansible-logging
ansible: raise an AnsibleFailedError even when parsing fails
Andrew Schoen [Wed, 19 Aug 2015 16:06:22 +0000 (11:06 -0500)]
ansible: raise an AnsibleFailedError even when parsing fails
If _handle_failures fails to parse the yaml in the ansible error log, it
can still give the contents of the failure log to an AnsibleFailedError
in it's unparsed form.
I also changed the log.exception to log.error when an ParserError occurs
as I think it was confusing some having that extra traceback in the log.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Zack Cerza [Wed, 19 Aug 2015 16:01:30 +0000 (10:01 -0600)]
Merge pull request #572 from ivotron/doc-enhancement-on-lab-components
Documentation on lab components
Ivo Jimenez [Wed, 19 Aug 2015 01:15:41 +0000 (18:15 -0700)]
Adds documentation about components of a Ceph Lab
Zack Cerza [Tue, 18 Aug 2015 16:29:18 +0000 (10:29 -0600)]
Merge pull request #599 from ceph/wip-archive-ansible-log
task.ansible: archive the ansible log even if parsing fails
Andrew Schoen [Tue, 18 Aug 2015 15:09:00 +0000 (10:09 -0500)]
task.ansible: archive the ansible log even if parsing fails
This fixes a bug where if we fail to parse the ansible log it is never
archived.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Ken Dreyer [Tue, 18 Aug 2015 14:50:47 +0000 (08:50 -0600)]
Merge pull request #598 from ceph/wip-ansible-log-perms
task.ansible: set the ansible_failure.yaml file mode to 0664
Andrew Schoen [Mon, 17 Aug 2015 21:06:13 +0000 (16:06 -0500)]
task.ansible: set the ansible_failure.yaml file mode to 0664
The failure log was originally created with mode 0600 which does not
allow apache to serve the file. This changes the mode to 0664 to match
the rest of the archived files from a teuthology job.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Gregory Farnum [Tue, 18 Aug 2015 11:21:21 +0000 (12:21 +0100)]
Merge pull request #554 from ceph/wip-9994
tasks/nfs: make sure NFS kernel server is started
Reviewed-by: Greg Farnum <gfarnum@redhat.com>
Zack Cerza [Mon, 17 Aug 2015 18:03:53 +0000 (12:03 -0600)]
Merge pull request #596 from ceph/wip-12668
refactor the install task to use packaging.GitbuilderProject
Dan Mick [Fri, 14 Aug 2015 16:05:43 +0000 (09:05 -0700)]
Merge pull request #597 from ceph/wip-fix-ansible-again
ansible: rename log to fail_log to not conflict with ansible.log
Reviewed-by: Dan Mick <dmick@redhat.com>
Andrew Schoen [Fri, 14 Aug 2015 14:10:32 +0000 (09:10 -0500)]
ansible: rename log to fail_log to not conflict with ansible.log
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Andrew Schoen [Wed, 12 Aug 2015 16:48:02 +0000 (11:48 -0500)]
task.install: use packing.GitbuilderProject to interface with gitbuilder
This refactors the install and upgrade methods in the install task to
use packing.GitbuilderProject for both the rpm and deb code paths.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Andrew Schoen [Tue, 11 Aug 2015 16:20:42 +0000 (11:20 -0500)]
packaging.GitbuilderProject: add a test to ensure base_url is correct for wheezy
remote.os.codename returns an empty string for debian which resulted in
the base_url missing the 'wheezy' part of it, this fixes that.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Andrew Schoen [Tue, 11 Aug 2015 15:32:49 +0000 (10:32 -0500)]
Adds a test for debian 7.1 to make sure wheezy is selected
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Alfredo Deza [Thu, 13 Aug 2015 17:52:52 +0000 (13:52 -0400)]
Merge pull request #595 from ceph/wip-12656
fix ansible error reporting
Andrew Schoen [Thu, 13 Aug 2015 14:32:45 +0000 (09:32 -0500)]
task.ansible: use shutil.move instead of os.rename to move a failure log
When moving a file between file systems of different types os.rename will
throw an exception, while shutil.move correctly handles that case.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Andrew Schoen [Thu, 13 Aug 2015 14:12:06 +0000 (09:12 -0500)]
task.ansible: catch failure log parse errors
If we fail to parse a failure log, fail silently and log the exception
so that we can inspect that log after the fact.
In this case a CommandFailedError will be raised.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Andrew Schoen [Thu, 6 Aug 2015 17:24:38 +0000 (12:24 -0500)]
Merge pull request #593 from ceph/wip-12605
SELinux: Rotate and archive audit.log
Zack Cerza [Wed, 5 Aug 2015 22:27:28 +0000 (16:27 -0600)]
Archive audit log
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Wed, 5 Aug 2015 22:11:16 +0000 (16:11 -0600)]
Rotate audit logs before checking denials
Signed-off-by: Zack Cerza <zack@redhat.com>
Andrew Schoen [Wed, 5 Aug 2015 19:32:56 +0000 (14:32 -0500)]
Merge pull request #587 from ceph/wip-silent-signal
orchestra/daemon: add 'silent' option to signal()
Andrew Schoen [Wed, 5 Aug 2015 17:43:08 +0000 (12:43 -0500)]
Merge pull request #590 from ceph/wip-remote-chcon
Add Remote.chcon()
Zack Cerza [Wed, 5 Aug 2015 17:42:41 +0000 (11:42 -0600)]
Merge pull request #589 from ceph/wip-better-ansible-failures
ansible: read the ansible failure log and raise a more meaningful error
Zack Cerza [Wed, 5 Aug 2015 16:06:10 +0000 (10:06 -0600)]
Use Remote.chcon() in task.internal.syslog()
Signed-off-by: Zack Cerza <zack@redhat.com>
Zack Cerza [Wed, 5 Aug 2015 16:05:17 +0000 (10:05 -0600)]
Add Remote.chcon()
Signed-off-by: Zack Cerza <zack@redhat.com>
Andrew Schoen [Tue, 4 Aug 2015 20:10:25 +0000 (15:10 -0500)]
ansible: read the ansible failure log and raise a more meaningful error
This works with a custom ansible callback plugin that intercepts task
failures and writes them to a log file (in yaml format) provided by the
ansible task as an environment variable. If the playbook fails, teuthology
reads that failure log and raises an AnsibleFailedError with the actual
output from the failed ansible task.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Zack Cerza [Tue, 4 Aug 2015 17:52:54 +0000 (11:52 -0600)]
Merge pull request #588 from ceph/wip-ansible-failures
install: ignore minor version when building a gitbuilder url
Andrew Schoen [Tue, 4 Aug 2015 17:12:49 +0000 (12:12 -0500)]
install: ignore minor version when building a gitbuilder url
We don't care about minor version in gitbuilder anymore.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Sage Weil [Tue, 4 Aug 2015 17:06:44 +0000 (13:06 -0400)]
orchestra/daemon: add 'silent' option to signal()
Default to False (not silent) to preserve existing behavior.
Signed-off-by: Sage Weil <sage@redhat.com>
Zack Cerza [Tue, 4 Aug 2015 15:35:30 +0000 (09:35 -0600)]
Merge pull request #586 from ceph/wip-no-minor-gitbuilder
packaging: ignore minor version numbers in gitbuilder urls
Andrew Schoen [Tue, 4 Aug 2015 14:24:07 +0000 (09:24 -0500)]
packaging: ignore minor version numbers in gitbuilder urls
We don't care about minor version numbers anymore for gitbuilder.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Zack Cerza [Mon, 3 Aug 2015 20:07:29 +0000 (14:07 -0600)]
Merge pull request #580 from ceph/wip-12220
Introduces teuthology.packaging.GitbuilderProject
Andrew Schoen [Fri, 31 Jul 2015 21:14:33 +0000 (16:14 -0500)]
use packaging.DEFAULT_OS_VERSION in teuthology.misc
No need to keep two copies of this data around
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Andrew Schoen [Fri, 31 Jul 2015 21:06:06 +0000 (16:06 -0500)]
remove the need for teuthology.misc in teuthology.packaging
It's best to get the package type from remote.os.package_type. This also
fixes a circular import that will allows me to import
teuthology.packaging into teuthology.misc
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Andrew Schoen [Thu, 30 Jul 2015 16:35:10 +0000 (11:35 -0500)]
GitbuilderProject: change uri to uri_reference
It is not technically a full uri, but a uri reference.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Andrew Schoen [Thu, 30 Jul 2015 15:11:16 +0000 (10:11 -0500)]
Better documentation for packaging.GitbuilderProject
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Andrew Schoen [Wed, 22 Jul 2015 22:15:56 +0000 (17:15 -0500)]
task.internal: use packaging.GitbuilderProject in internal.check_packages
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Andrew Schoen [Wed, 22 Jul 2015 21:03:11 +0000 (16:03 -0500)]
Introduces packaging.GitbuilderProject
This is a new class that can be used to interface with gitbuilder to get
information about a project. Currently it's used primarily for version
checks and constructing the correct base url for a projects location on
gitbuilder.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Andrew Schoen [Mon, 6 Jul 2015 21:57:51 +0000 (16:57 -0500)]
task.install: make _get_baseurlinfo_and_dist use remote.os
This also adds some tests and a helper function.
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
Zack Cerza [Mon, 3 Aug 2015 16:39:21 +0000 (10:39 -0600)]
Merge pull request #584 from ceph/wip-rest-bench
task/install: rest-bench has been removed
Sage Weil [Mon, 3 Aug 2015 16:34:50 +0000 (12:34 -0400)]
task/install: rest-bench has been removed
Removed this tool from upstream. Don't install (or remove) it any more!
Signed-off-by: Sage Weil <sage@redhat.com>
Zack Cerza [Fri, 31 Jul 2015 19:55:08 +0000 (13:55 -0600)]
Merge pull request #500 from ceph/wip-kernel-overrides
Support overrides: in the kernel task
Zack Cerza [Fri, 31 Jul 2015 19:21:10 +0000 (13:21 -0600)]
Merge pull request #519 from ceph/wip-f22
support fedora 22
Sage Weil [Fri, 31 Jul 2015 15:39:31 +0000 (11:39 -0400)]
provision: specify ubuntu in __init__
Signed-off-by: Sage Weil <sage@redhat.com>
Ilya Dryomov [Thu, 7 May 2015 10:55:14 +0000 (13:55 +0300)]
kernel: support overrides
This is needed for the new unmap subsuite of krbd suite.
normalize_config() config should actually copy config snippets into
new_config, otherwise popping keys from one role's snippet would also
affect other roles, potentially resulting in overrides: overriding more
than it should have.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Ilya Dryomov [Thu, 30 Jul 2015 17:42:25 +0000 (20:42 +0300)]
kernel: amend normalize_config() logic
In the case where we stamp out a config for all roles, don't take
a shortcut and stamp it out once per remote. Instead, do it for each
role, so that the number of items in the resulting dict in this case is
equal to the number of roles and not to the number of remotes. This is
a necessary prerequisite for override support - otherwise something
like
kernel:
client:
branch: testing
override:
kernel:
branch: wip-foobar
won't be deep_merged()'ed correctly if the role:remote mapping looks
like
remote1: mon.a, osd.0, osd.1, osd2
remote2: mds.a, client.0
because we'd have
mon.a: branch: wip-foobar
mds.a: branch: wip-foobar
on the override side and
client.0: branch: testing
on the original config side. Since those three keys don't overlap,
we'd end up with testing kernel on client.0, which is incorrect.
This doesn't result in any functional changes - validate_config()
will reduce such config as usual.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Ilya Dryomov [Thu, 7 May 2015 10:30:09 +0000 (13:30 +0300)]
kernel: separate version keys into a list, use constants
Use constants for config snippet and timeout defaults.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>