]> git.apps.os.sepia.ceph.com Git - teuthology.git/log
teuthology.git
10 years agoMerge pull request #488 from ceph/wip-11552
Zack Cerza [Tue, 12 May 2015 15:37:04 +0000 (09:37 -0600)]
Merge pull request #488 from ceph/wip-11552

Install scratch built kernels from koji

10 years agoFix copypasta
Zack Cerza [Mon, 11 May 2015 21:27:38 +0000 (15:27 -0600)]
Fix copypasta

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoAdds support for installing kernels from koji task results. 488/head
Andrew Schoen [Mon, 11 May 2015 15:47:00 +0000 (10:47 -0500)]
Adds support for installing kernels from koji task results.

The kernel task now takes a task_id argument that can be used to
download and install a kernel from task results.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
10 years agoAdds methods to extract info on a package from koji task results.
Andrew Schoen [Mon, 11 May 2015 15:26:22 +0000 (10:26 -0500)]
Adds methods to extract info on a package from koji task results.

Once you get the results of a koji task you can use these methods to
look for and extract information about a given package.  If the package
is found it will return information about how to download and consume
the rpm generated by the task.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
10 years agoAdds packaging.get_koji_task_result, to query kojihub for a tasks result.
Andrew Schoen [Thu, 7 May 2015 21:18:25 +0000 (16:18 -0500)]
Adds packaging.get_koji_task_result, to query kojihub for a tasks result.

We need this because we want to be able to use scratch builds, which do
not generate a build ID. We can download the rpms we need by
retrieving the results of a task then looking at the list of rpms
it created.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
10 years agoMerge pull request #487 from ceph/wip-ansible
Andrew Schoen [Mon, 11 May 2015 18:27:06 +0000 (13:27 -0500)]
Merge pull request #487 from ceph/wip-ansible

Ansible task

10 years agoClarify Task.filter_hosts() docstring 487/head
Zack Cerza [Mon, 11 May 2015 15:57:05 +0000 (09:57 -0600)]
Clarify Task.filter_hosts() docstring

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoAnsible task tests
Zack Cerza [Thu, 7 May 2015 20:46:35 +0000 (14:46 -0600)]
Ansible task tests

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoAnsible task
Zack Cerza [Tue, 21 Apr 2015 16:45:25 +0000 (10:45 -0600)]
Ansible task

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoSort hostnames in unit test
Zack Cerza [Fri, 8 May 2015 15:53:35 +0000 (09:53 -0600)]
Sort hostnames in unit test

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMerge pull request #486 from ceph/wip-schedule
Zack Cerza [Fri, 8 May 2015 15:43:46 +0000 (09:43 -0600)]
Merge pull request #486 from ceph/wip-schedule

Wip schedule -- with fix for 11533

10 years agomatrix: fix Concat.index -- should be (item, {concat_items}) 486/head
Samuel Just [Thu, 7 May 2015 19:56:53 +0000 (12:56 -0700)]
matrix: fix Concat.index -- should be (item, {concat_items})

Previously, it just returned {concat_items} which isn't right.

Signed-off-by: Samuel Just <sjust@redhat.com>
10 years agoMerge pull request #479 from zhouyuan/config_max_reserve
Zack Cerza [Thu, 7 May 2015 21:28:29 +0000 (15:28 -0600)]
Merge pull request #479 from zhouyuan/config_max_reserve

Make the number machines to be reserved configurable

10 years agosuite: add --subset option to allow us to schedule subsets of a suite
Samuel Just [Fri, 27 Mar 2015 17:20:28 +0000 (10:20 -0700)]
suite: add --subset option to allow us to schedule subsets of a suite

First, the old logic tended to do the following:

[0,1,2,3] x [a,b,c,d] -> [(0, a), (0, b), (0, c), ...]

which is a bummer since if we only schedule the first 1/4,
we only get combinations with 0.

Instead, we want:

[0,1,2,3] x [a,b,c,d] -> [(0, a), (1, b), (2, c), ...]

Second, the old logic tended to do the following:

[0,1,2,3] + [a,b,c,d] -> [0,1,2,3,a,b,c,d]

Which is a bummer since we need to all the way through the first
set before we get examples from the second.

Instead, we want:

[0,1,2,3] + [a,b,c,d] -> [0,a,1,b,2,c,3,d]

Lastly, if we have:

[0,1,2] + ([0,1,2],[a,b,c])

and want to schedule 1/3 of the suite, we don't want to
just schedule 0, we probably want to schedule all of
[0,1,2] and 1/3 of the second combo.  Thus, we want to
detect such cases and multiply out [0,1,2] into
[0,1,2,0,1,2,0,1,2].

Signed-off-by: Samuel Just <sam.just@inktank.com>
(cherry picked from commit 9a57759ec3df7a8b9490920538214ac189418c49)

10 years agotest_suite: add test for bug 11533
Samuel Just [Thu, 7 May 2015 19:18:31 +0000 (12:18 -0700)]
test_suite: add test for bug 11533

Signed-off-by: Samuel Just <sjust@redhat.com>
10 years agoMerge pull request #483 from SUSE/wip-suse-bootsrap
Andrew Schoen [Tue, 5 May 2015 20:15:23 +0000 (15:15 -0500)]
Merge pull request #483 from SUSE/wip-suse-bootsrap

Added openSUSE and SLES conditions in bootstrap.

10 years agoAdded openSUSE and SLES conditions in bootstrap. 483/head
ksharma [Tue, 5 May 2015 14:35:38 +0000 (16:35 +0200)]
Added openSUSE and SLES conditions in bootstrap.

Signed-off-by: Kapil Sharma ksharma@suse.com
10 years agoRevert "suite: add --subset option to allow us to schedule subsets of a suite"
Zack Cerza [Tue, 5 May 2015 01:39:40 +0000 (19:39 -0600)]
Revert "suite: add --subset option to allow us to schedule subsets of a suite"

This reverts commit 9a57759ec3df7a8b9490920538214ac189418c49.

See http://tracker.ceph.com/issues/11533

10 years agoMerge pull request #340 from ceph/wip-schedule
Zack Cerza [Mon, 4 May 2015 22:28:13 +0000 (16:28 -0600)]
Merge pull request #340 from ceph/wip-schedule

suite: add --subset option to allow us to schedule subsets of a suit...

10 years agosuite: add --subset option to allow us to schedule subsets of a suite 340/head
Samuel Just [Fri, 27 Mar 2015 17:20:28 +0000 (10:20 -0700)]
suite: add --subset option to allow us to schedule subsets of a suite

First, the old logic tended to do the following:

[0,1,2,3] x [a,b,c,d] -> [(0, a), (0, b), (0, c), ...]

which is a bummer since if we only schedule the first 1/4,
we only get combinations with 0.

Instead, we want:

[0,1,2,3] x [a,b,c,d] -> [(0, a), (1, b), (2, c), ...]

Second, the old logic tended to do the following:

[0,1,2,3] + [a,b,c,d] -> [0,1,2,3,a,b,c,d]

Which is a bummer since we need to all the way through the first
set before we get examples from the second.

Instead, we want:

[0,1,2,3] + [a,b,c,d] -> [0,a,1,b,2,c,3,d]

Lastly, if we have:

[0,1,2] + ([0,1,2],[a,b,c])

and want to schedule 1/3 of the suite, we don't want to
just schedule 0, we probably want to schedule all of
[0,1,2] and 1/3 of the second combo.  Thus, we want to
detect such cases and multiply out [0,1,2] into
[0,1,2,0,1,2,0,1,2].

Signed-off-by: Samuel Just <sam.just@inktank.com>
10 years agoMerge pull request #481 from ceph/wip-11429
Samuel Just [Mon, 4 May 2015 21:19:27 +0000 (14:19 -0700)]
Merge pull request #481 from ceph/wip-11429

Wip 11429

Reviewed-by: Andrew Schoen <andrew.schoen@gmail.com>
10 years agotask: add loop 481/head
Samuel Just [Mon, 4 May 2015 17:41:21 +0000 (10:41 -0700)]
task: add loop

10 years agotask: add full_sequential
Samuel Just [Mon, 4 May 2015 16:38:18 +0000 (09:38 -0700)]
task: add full_sequential

10 years agoMake the number machines to be reserved configurable 479/head
Yuan Zhou [Thu, 30 Apr 2015 06:53:14 +0000 (14:53 +0800)]
Make the number machines to be reserved configurable

Signed-off-by: Yuan Zhou <yuan.zhou@intel.com>
10 years agoMerge pull request #478 from ceph/nuke-updates
Zack Cerza [Thu, 30 Apr 2015 16:37:37 +0000 (10:37 -0600)]
Merge pull request #478 from ceph/nuke-updates

Don't assert status==up when using internal.check_lock.

10 years agoDon't assert status==up when using internal.check_lock. 478/head
Andrew Schoen [Wed, 29 Apr 2015 15:43:52 +0000 (11:43 -0400)]
Don't assert status==up when using internal.check_lock.

Add a check_up kwarg to check_lock so that we can optionally skip
checking to see if a node is up before nuking it. This still allows
scheduled teuthology jobs to make sure they didn't lock a downed node.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
10 years agoCall add_remotes() before connect()
Zack Cerza [Tue, 28 Apr 2015 20:48:16 +0000 (14:48 -0600)]
Call add_remotes() before connect()

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoRun context manager exit if enter fails
Zack Cerza [Tue, 28 Apr 2015 19:26:27 +0000 (13:26 -0600)]
Run context manager exit if enter fails

In run_tasks(), we were waiting until a context manager's __enter__()
finished before adding it to the stack. As a result, if __enter__()
failed, __exit__() would never be called.

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMerge pull request #477 from ceph/wip-11491
Zack Cerza [Tue, 28 Apr 2015 16:27:45 +0000 (10:27 -0600)]
Merge pull request #477 from ceph/wip-11491

include E9 errors for flake8 runs in tox.ini

10 years agoinclude E9 errors for flake8 runs in tox.ini 477/head
Alfredo Deza [Tue, 28 Apr 2015 16:16:46 +0000 (12:16 -0400)]
include E9 errors for flake8 runs in tox.ini

Signed-off-by: Alfredo Deza <alfredo.deza@inktank.com>
10 years agoFix broken string
Zack Cerza [Tue, 28 Apr 2015 01:12:37 +0000 (19:12 -0600)]
Fix broken string

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMerge pull request #476 from ceph/wip-11474
Andrew Schoen [Mon, 27 Apr 2015 21:58:34 +0000 (16:58 -0500)]
Merge pull request #476 from ceph/wip-11474

Log more output in need_to_install_distro()

10 years agoAdd fetch_repo() and use it
Zack Cerza [Wed, 22 Apr 2015 22:13:58 +0000 (16:13 -0600)]
Add fetch_repo() and use it

Abstracts the copy-pasta codepath used by fetch_qa_suite() and
fetch_teuthology() into a function usable by both, and also usable to
fetch arbitrary repos. Provides branch selection, optional locking and
retry, and an optional bootstrap callback.

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoFix CommandFailedError when node is None
Zack Cerza [Wed, 22 Apr 2015 21:50:27 +0000 (15:50 -0600)]
Fix CommandFailedError when node is None

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoFix filter_hosts() when no filters are specified
Zack Cerza [Mon, 27 Apr 2015 21:54:24 +0000 (15:54 -0600)]
Fix filter_hosts() when no filters are specified

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMerge pull request #475 from ceph/wip-newtasks
Andrew Schoen [Mon, 27 Apr 2015 20:30:02 +0000 (15:30 -0500)]
Merge pull request #475 from ceph/wip-newtasks

New-style tasks

10 years agoAdd unit tests for teuthology.task.Task 475/head
Zack Cerza [Thu, 23 Apr 2015 21:43:30 +0000 (15:43 -0600)]
Add unit tests for teuthology.task.Task

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoAdd teuthology.task.Task
Zack Cerza [Tue, 21 Apr 2015 16:44:54 +0000 (10:44 -0600)]
Add teuthology.task.Task

This is intended to be a base class providing a more modern, featured,
tested foundation for tasks. It has built-in overrides, host selection,
and optional skipping of the teardown stage.

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoneed_to_install_distro(): log more output 476/head
Zack Cerza [Mon, 27 Apr 2015 19:10:07 +0000 (13:10 -0600)]
need_to_install_distro(): log more output

This was another case where stdout/stderr was being supressed
needlessly. We need a bit of it for debugging purposes.

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoClean up linter issues in need_to_install_distro()
Zack Cerza [Mon, 27 Apr 2015 19:02:44 +0000 (13:02 -0600)]
Clean up linter issues in need_to_install_distro()

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMerge pull request #474 from ceph/wip-11466
Andrew Schoen [Fri, 24 Apr 2015 16:04:05 +0000 (11:04 -0500)]
Merge pull request #474 from ceph/wip-11466

Fix branch selection logic

10 years agoFix branch selection logic 474/head
Zack Cerza [Fri, 24 Apr 2015 16:01:30 +0000 (10:01 -0600)]
Fix branch selection logic

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMerge pull request #469 from zhouyuan/add_missing_files
Zack Cerza [Thu, 23 Apr 2015 23:31:32 +0000 (17:31 -0600)]
Merge pull request #469 from zhouyuan/add_missing_files

Adding missing files in setup.py

10 years agoMerge pull request #473 from ceph/wip-rate-limit
Dan Mick [Thu, 23 Apr 2015 23:14:05 +0000 (16:14 -0700)]
Merge pull request #473 from ceph/wip-rate-limit

Don't call get_branch_info() for master branches

Reviewed-by: Dan Mick <dmick@redhat.com>
10 years agoMerge pull request #471 from dmick/master
Zack Cerza [Thu, 23 Apr 2015 23:11:10 +0000 (17:11 -0600)]
Merge pull request #471 from dmick/master

ssh_keys: save each remote's backup file name separately in a dict for u...

10 years agoDon't call get_branch_info() for master branches 473/head
Zack Cerza [Thu, 23 Apr 2015 23:05:31 +0000 (17:05 -0600)]
Don't call get_branch_info() for master branches

We're hitting the github API rate limit; if master doesn't exist we lose
anyway

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agossh_keys: save each remote's backup file name separately in a dict for undo 467/head 471/head
Dan Mick [Thu, 23 Apr 2015 21:34:56 +0000 (14:34 -0700)]
ssh_keys: save each remote's backup file name separately in a dict for undo

Signed-off-by: Dan Mick <dan.mick@redhat.com>
10 years agoMerge pull request #470 from ceph/wip-remote
Andrew Schoen [Thu, 23 Apr 2015 19:49:30 +0000 (14:49 -0500)]
Merge pull request #470 from ceph/wip-remote

Add timeouts to Remote connection functions

10 years agoSplit internal.connect() into two subtasks 470/head
Zack Cerza [Thu, 23 Apr 2015 16:51:54 +0000 (10:51 -0600)]
Split internal.connect() into two subtasks

One creates the ctx.cluster object; the other initiates connections

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoPass timeout value down to connection.connect()
Zack Cerza [Thu, 23 Apr 2015 16:50:58 +0000 (10:50 -0600)]
Pass timeout value down to connection.connect()

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoAdd timeout arg to connect()
Zack Cerza [Thu, 23 Apr 2015 16:21:46 +0000 (10:21 -0600)]
Add timeout arg to connect()

Default behavior is unchanged

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoDon't call Remote.connect() in Remote.__init__()
Zack Cerza [Wed, 15 Apr 2015 18:31:08 +0000 (12:31 -0600)]
Don't call Remote.connect() in Remote.__init__()

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMake Remote.reconnect() retry optionally
Zack Cerza [Wed, 15 Apr 2015 18:15:44 +0000 (12:15 -0600)]
Make Remote.reconnect() retry optionally

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoProperly close connection on object deletion
Zack Cerza [Wed, 15 Apr 2015 16:58:04 +0000 (10:58 -0600)]
Properly close connection on object deletion

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoAdding missing files in setup.py 469/head
Yuan Zhou [Thu, 23 Apr 2015 01:37:50 +0000 (09:37 +0800)]
Adding missing files in setup.py

Some non-python scripts are missing when install.

Signed-off-by: Yuan Zhou <yuan.zhou@intel.com>
10 years agoMerge pull request #468 from dmick/wip-ssh-keys
Zack Cerza [Tue, 21 Apr 2015 17:56:38 +0000 (11:56 -0600)]
Merge pull request #468 from dmick/wip-ssh-keys

ssh_keys: stop using sed, use saved backup of authorized_keys instead

10 years agossh_keys: stop using sed, use saved backup of authorized_keys instead 468/head
Dan Mick [Tue, 21 Apr 2015 05:18:01 +0000 (22:18 -0700)]
ssh_keys: stop using sed, use saved backup of authorized_keys instead

Signed-off-by: Dan Mick <dan.mick@redhat.com>
10 years agoMerge pull request #466 from ceph/wip-11426
Andrew Schoen [Mon, 20 Apr 2015 17:38:29 +0000 (12:38 -0500)]
Merge pull request #466 from ceph/wip-11426

Log stderr in get_latest_image_version_deb()

10 years agoLog stderr in get_latest_image_version_deb() 466/head
Zack Cerza [Mon, 20 Apr 2015 17:25:24 +0000 (11:25 -0600)]
Log stderr in get_latest_image_version_deb()

It was being suppressed needlessly. Also fix lots of linter issues.

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMerge pull request #465 from ceph/kernel-koji-docs
Zack Cerza [Mon, 20 Apr 2015 16:41:34 +0000 (10:41 -0600)]
Merge pull request #465 from ceph/kernel-koji-docs

Docs for using koji in the kernel task.

10 years agoDocs for using koji in the kernel task. 465/head
Andrew Schoen [Mon, 20 Apr 2015 16:19:11 +0000 (11:19 -0500)]
Docs for using koji in the kernel task.

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
10 years agoMerge pull request #464 from ceph/wip-git.ceph.com
Zack Cerza [Thu, 16 Apr 2015 18:35:48 +0000 (12:35 -0600)]
Merge pull request #464 from ceph/wip-git.ceph.com

git.ceph.com

10 years agogit.ceph.com 464/head
Sage Weil [Thu, 16 Apr 2015 15:39:25 +0000 (08:39 -0700)]
git.ceph.com

Signed-off-by: Sage Weil <sage@redhat.com>
10 years agoMerge pull request #462 from ceph/wip-ssh-keys
Dan Mick [Wed, 15 Apr 2015 17:30:46 +0000 (10:30 -0700)]
Merge pull request #462 from ceph/wip-ssh-keys

When modifying authorized_keys, store a backup

10 years agoWhen modifying authorized_keys, store a backup 462/head
Zack Cerza [Mon, 13 Apr 2015 23:20:19 +0000 (17:20 -0600)]
When modifying authorized_keys, store a backup

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMerge pull request #460 from zhouyuan/mkdir_p
Zack Cerza [Tue, 14 Apr 2015 17:38:15 +0000 (11:38 -0600)]
Merge pull request #460 from zhouyuan/mkdir_p

Make parent directories as needed

10 years agoMerge pull request #463 from ceph/wip-suite-tests
Andrew Schoen [Tue, 14 Apr 2015 17:27:07 +0000 (12:27 -0500)]
Merge pull request #463 from ceph/wip-suite-tests

Basic teuthology.suite.build_matrix() unit tests

10 years agoMake test_gitbuilder_url() use the local config 463/head
Zack Cerza [Tue, 14 Apr 2015 15:30:14 +0000 (09:30 -0600)]
Make test_gitbuilder_url() use the local config

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoAdd unit tests for suite.build_matrix()
Zack Cerza [Mon, 13 Apr 2015 22:18:13 +0000 (16:18 -0600)]
Add unit tests for suite.build_matrix()

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoParametrize build_matrix() for unit tests
Zack Cerza [Mon, 13 Apr 2015 22:17:40 +0000 (16:17 -0600)]
Parametrize build_matrix() for unit tests

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoAlways call unlock_one() with the user arg
Zack Cerza [Fri, 10 Apr 2015 16:30:53 +0000 (10:30 -0600)]
Always call unlock_one() with the user arg

I'd seen a few cases where vps creation would fail, and then unlocking
failed because we were incorrectly guessing the owner of the lock.

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoSkip passing os_type/version when destroying
Zack Cerza [Thu, 9 Apr 2015 21:42:13 +0000 (15:42 -0600)]
Skip passing os_type/version when destroying

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMerge pull request #459 from ceph/wip-downburst
Andrew Schoen [Thu, 9 Apr 2015 20:50:10 +0000 (15:50 -0500)]
Merge pull request #459 from ceph/wip-downburst

Refactor downburst provisioning

10 years agoAdd basic unit tests for Downburst 459/head
Zack Cerza [Thu, 9 Apr 2015 16:33:26 +0000 (10:33 -0600)]
Add basic unit tests for Downburst

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMake *_if_vm() more unit-testable
Zack Cerza [Thu, 9 Apr 2015 16:30:45 +0000 (10:30 -0600)]
Make *_if_vm() more unit-testable

Allow passing in a custom Downburst instance. Also clean up several
linter issues.

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoUse provision.downburst_executable()
Zack Cerza [Fri, 27 Feb 2015 20:40:07 +0000 (13:40 -0700)]
Use provision.downburst_executable()

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMerge pull request #461 from ceph/wip-update-inventory
Andrew Schoen [Tue, 7 Apr 2015 22:54:07 +0000 (17:54 -0500)]
Merge pull request #461 from ceph/wip-update-inventory

Add teuthology-update-inventory command

10 years agoAdd teuthology-update-inventory command 461/head
Zack Cerza [Tue, 7 Apr 2015 16:54:28 +0000 (10:54 -0600)]
Add teuthology-update-inventory command

This ought to also replace teuthology-updatekeys, but might not be as
fast yet.

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMove downburst code into new Downburst class
Zack Cerza [Fri, 13 Feb 2015 17:41:36 +0000 (10:41 -0700)]
Move downburst code into new Downburst class

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMake parent directories as needed 460/head
Yuan Zhou [Tue, 7 Apr 2015 04:35:38 +0000 (12:35 +0800)]
Make parent directories as needed

with '-p' when mkdir

Signed-off-by: Yuan Zhou <yuan.zhou@intel.com>
10 years agoClarify error logging in destroy_if_vm()
Zack Cerza [Fri, 13 Feb 2015 03:20:45 +0000 (20:20 -0700)]
Clarify error logging in destroy_if_vm()

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoBe smarter about logging in create_if_vm()
Zack Cerza [Thu, 12 Feb 2015 21:57:57 +0000 (14:57 -0700)]
Be smarter about logging in create_if_vm()

Also return False on failure.

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoStrip ssh-keyscan output before logging it
Zack Cerza [Thu, 12 Feb 2015 21:54:32 +0000 (14:54 -0700)]
Strip ssh-keyscan output before logging it

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoBe more tolerant of git connection errors
Zack Cerza [Fri, 3 Apr 2015 03:41:43 +0000 (21:41 -0600)]
Be more tolerant of git connection errors

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoBe slightly smarter when looking for stale locks
Zack Cerza [Wed, 1 Apr 2015 23:30:48 +0000 (17:30 -0600)]
Be slightly smarter when looking for stale locks

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMerge pull request #458 from ceph/wip-11248
Zack Cerza [Fri, 27 Mar 2015 18:58:53 +0000 (12:58 -0600)]
Merge pull request #458 from ceph/wip-11248

Disable version check when ceph-deploy is used to install ceph.

10 years agoDisable version check when ceph-deploy is used to install ceph. 458/head
Andrew Schoen [Fri, 27 Mar 2015 17:43:04 +0000 (10:43 -0700)]
Disable version check when ceph-deploy is used to install ceph.

See: http://tracker.ceph.com/issues/11248

Signed-off-by: Andrew Schoen <aschoen@redhat.com>
10 years agoOn BranchNotFoundError, mark job as dead
Zack Cerza [Fri, 27 Mar 2015 17:23:16 +0000 (11:23 -0600)]
On BranchNotFoundError, mark job as dead

We had been simply deleting jobs from paddles if the teuthology branch
was not found. Instead, mark them as dead with a useful failure_reason.

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMerge pull request #457 from ceph/wip-lttng-valgrind
Andrew Schoen [Thu, 26 Mar 2015 22:41:31 +0000 (15:41 -0700)]
Merge pull request #457 from ceph/wip-lttng-valgrind

valgrind: ignore lttng warning on un-cleaned-up state

10 years agovalgrind: ignore lttng warning on un-cleaned-up state 457/head
Greg Farnum [Thu, 26 Mar 2015 21:13:45 +0000 (14:13 -0700)]
valgrind: ignore lttng warning on un-cleaned-up state

This appears to have come from dca722ec7b2a7fc9214844ec92310074b5cb2faa,
which merged in support for use of lttng ust functions on fork. Valgrind
started warning on a Leak_StillReachable with lttng_ust_init in
the function call trace.

Hopefully this is specific enough that we don't mask out any other
lttng errors.

Fixes: #11247
Signed-off-by: Greg Farnum <gfarnum@redhat.com>
10 years agoMerge pull request #456 from ceph/wip-11243
Zack Cerza [Thu, 26 Mar 2015 20:16:22 +0000 (13:16 -0700)]
Merge pull request #456 from ceph/wip-11243

Verify version for other projects besides ceph in the install task

10 years agoVerify version for other projects besides ceph in the install task 456/head
Andrew Schoen [Thu, 26 Mar 2015 16:37:03 +0000 (09:37 -0700)]
Verify version for other projects besides ceph in the install task

Fixes: http://tracker.ceph.com/issues/11243
Signed-off-by: Andrew Schoen <aschoen@redhat.com>
10 years agoAvoid double-unlocking with unlock_on_failure
Zack Cerza [Tue, 24 Mar 2015 23:13:35 +0000 (16:13 -0700)]
Avoid double-unlocking with unlock_on_failure

If both unlock_on_failure and nuke-on-error are set, don't unlock now
because we're just going to nuke (and unlock) later.

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoMerge pull request #451 from ceph/verify-version
Zack Cerza [Mon, 23 Mar 2015 19:05:41 +0000 (13:05 -0600)]
Merge pull request #451 from ceph/verify-version

Verify that the correct ceph version is installed on updates and install

10 years agoMerge pull request #454 from ktdreyer/wip-comment-libvirt-bootstrap
Zack Cerza [Mon, 23 Mar 2015 15:43:30 +0000 (09:43 -0600)]
Merge pull request #454 from ktdreyer/wip-comment-libvirt-bootstrap

bootstrap: explain the system python-libvirt pkg

10 years agobootstrap: explain the system python-libvirt pkg 454/head
Ken Dreyer [Mon, 23 Mar 2015 15:30:26 +0000 (09:30 -0600)]
bootstrap: explain the system python-libvirt pkg

In #sepia today, Zack explained the reasoning for the python-libvirt
package on the Debian family of distros.

10 years agoMerge pull request #453 from ceph/wip-11170
Andrew Schoen [Thu, 19 Mar 2015 21:19:55 +0000 (16:19 -0500)]
Merge pull request #453 from ceph/wip-11170

Allow running teuthology-lock --update on broken vps nodes

10 years agoTweak ssh-keyscan timing 453/head
Zack Cerza [Thu, 19 Mar 2015 20:58:26 +0000 (14:58 -0600)]
Tweak ssh-keyscan timing

Make the call itself timeout after 1s instead of 5s. Make update_lock()
try 15 times, sleeping for 1s between attempts. A full timeout of
update_lock() should take ~30s.

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoUse safe_while instead of an unsafe loop
Zack Cerza [Thu, 19 Mar 2015 20:32:58 +0000 (14:32 -0600)]
Use safe_while instead of an unsafe loop

If ssh-keyscan fails, we would hang forever. Stop that nonsense, and
give up eventually, print a warning, and move on.

Signed-off-by: Zack Cerza <zack@redhat.com>
10 years agoFix error formatting when 'action' is passed
Zack Cerza [Thu, 19 Mar 2015 20:32:06 +0000 (14:32 -0600)]
Fix error formatting when 'action' is passed

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