Activity
From 09/15/2016 to 10/14/2016
10/14/2016
- 10:16 PM Bug #17581: teuthology-kill failing when run as myself
- This appears to have been "the ssh-agent on the desktop was not loaded with the private key", for whatever reason
- 10:08 PM Bug #17581 (Rejected): teuthology-kill failing when run as myself
- This is all me and my box; somehow my ssh-agent has stopped functioning properly. I'm not sure when or how that happe...
- 09:54 PM Bug #17581: teuthology-kill failing when run as myself
- Dan asked if could ssh in as ubuntu user from myself. Short answer: no....
- 09:53 PM Bug #17581: teuthology-kill failing when run as myself
- I think this is probably a teuthology host config issue but am unsure which piece of infrastructure code is responsib...
- 09:52 PM Bug #17581 (Rejected): teuthology-kill failing when run as myself
- teuthology-kill fails because teuthology-nuke can't reach remote nodes....
10/13/2016
- 02:27 PM Revision 5137d3c6 (teuthology): ceph.conf: debug monc = 1
- Signed-off-by: Sage Weil <sage@redhat.com>
10/11/2016
- 08:43 PM Revision e365144c (teuthology): Add exit module
- This module contains a helper to add functions to be run when a given
signal is received. These signal handlers ought... - 08:10 PM Revision 97086c02 (teuthology): orchestra.connect: do not override user/host from ~/.ssh/config
- I don't know why this was ever the case, but ignoring user@host
from the targets list just seems like a bad idea, and...
10/08/2016
- 05:45 AM Bug #17125 (In Progress): teuthology-nuke fails when not running as ubuntu user
- 05:44 AM Bug #17125: teuthology-nuke fails when not running as ubuntu user
- Yeah. the more I think about it, the more I believe that the entire concept of overriding user or hostname is just ma...
- 05:20 AM Bug #17125: teuthology-nuke fails when not running as ubuntu user
- ...and the immediate failure is in mkdir /home/ubuntu/cephtest.
I think the right answer is to ignore User in the ... - 05:19 AM Bug #17125: teuthology-nuke fails when not running as ubuntu user
- Yes, that made the difference, and moreover, vps is a bad test vehicle, because I don't get an account made on the vp...
- 05:14 AM Bug #17125: teuthology-nuke fails when not running as ubuntu user
- ah...because the target name is ubuntu@<fqdn>, of course. Adding the full domain to my config file.
- 04:46 AM Bug #17125: teuthology-nuke fails when not running as ubuntu user
- So, looking into this, I'm confused. I have a config with a specific vpm set to User dmick and the rest of them set ...
10/07/2016
- 09:41 PM Feature #10945 (Resolved): Enable teuthology to re-run only failed jobs
- 07:42 PM Revision e689517d (teuthology): Ensure conserver/ipmitool processes die
- When the parent teuthology process dies unexpectedly, the process
logging console output sticks around. Use the atexi...
10/06/2016
- 10:02 PM Revision 1d3e5ed8 (teuthology): Merge pull request #963 from ceph/wip-rerun
- Add rerun capability to teuthology-suite
Reviewed-by: Dan Mick <dmick@redhat.com> - 09:12 PM Revision 503626a9 (teuthology): suite: Add rerun capability
- Using a new --rerun argument for teuthology-suite, optionally paired
with --rerun-statuses, we can now ask teuthology... - 06:03 PM Feature #10945 (In Progress): Enable teuthology to re-run only failed jobs
- PR open, would love feedback:
https://github.com/ceph/teuthology/pull/963 - 05:44 PM Revision ba13ca41 (teuthology): Add ResultsReporter.get_run()
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 04:50 PM Revision 2ffb0058 (teuthology): suite: parse filter args into lists sooner
- Signed-off-by: Zack Cerza <zack@redhat.com>
10/05/2016
- 05:27 PM Feature #10945: Enable teuthology to re-run only failed jobs
- 03:11 PM Bug #17351: stale dirs left at the end of the nuke
- This is probably a result of a previous run hitting a permission isue during nuke. http://tracker.ceph.com/issues/17125
- 03:10 PM Bug #17125: teuthology-nuke fails when not running as ubuntu user
- This is still a (frequently-hit) problem
10/04/2016
- 10:12 PM Feature #17498 (New): Teutology should not be using git builder to get kernel packages.
- Teuthology should be changed to use shaman/jenkins builds for kernel packages. As a first pass, I will try to find a...
- 08:32 PM Revision 27550b56 (teuthology): Merge pull request #951 from ceph/wip-shaman-2
- Support fetching packages from shaman instead of gitbuilder
Reviewed-by: Dan Mick <dmick@redhat.com> - 07:44 PM Revision be857fbe (teuthology): ceph.conf: debug auth on mon
- Tracking down an intermitted EPERM from mon... I think related to
rotating keys.
See http://tracker.ceph.com/issues/... - 03:14 PM Bug #17391 (Resolved): teuthology-nuke should lazy-unmount fuse mounts
- 03:01 PM Revision d18e36be (teuthology): More properly test ref precedence for Shaman
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 03:01 PM Revision 5c055b26 (teuthology): ShamanProject: correct ref precedence
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 03:01 PM Revision 5d1e0045 (teuthology): GitbuilderProject: remove 'v' from version values
- Some version values have a leading 'v', e.g.:
http://gitbuilder.ceph.com/ceph-rpm-centos7-x86_64-basic/ref/v10.2.0/ve... - 03:01 PM Revision e87fe10c (teuthology): Move much of TestGitbuilderProject into base class
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 03:01 PM Revision d702281c (teuthology): Add ShamanProject
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 03:01 PM Revision d09d6355 (teuthology): Add and use get_builder_project()
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 03:01 PM Revision 0986c327 (teuthology): task.install: use get_builder_project()
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 03:01 PM Revision c9af2f45 (teuthology): suite tests: use GitbuilderProject and not ShamanProject
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 03:01 PM Revision 8eeafc92 (teuthology): run: Allow a job to override config.use_shaman
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 03:01 PM Revision 865df5e4 (teuthology): task.internal: use get_builder_project()
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 03:01 PM Revision 4d2c0b6f (teuthology): get_package_version(): return V-R for rpm
- Even though strictly speaking "version" implies just the %{VERSION}
field, that is hardly ever useful on its own.
Si... - 03:01 PM Revision c2ba13e4 (teuthology): GitbuilderProject: don't mangle rpm versions
- Up until now, given a package like:
"ceph-test-11.0.0-2135.g8403c95.x86_64", we were returning "11.0.0",
which doesn'... - 03:01 PM Revision 03e5bc35 (teuthology): task.install: use builder.codename for deb repos
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 03:01 PM Revision 0edfc585 (teuthology): GitbuilderProject: always set codename
- When a GitbuilderProject (and also ShamanProject) was being init-ed from
a Remote object, we weren't setting the code... - 03:01 PM Revision 9a821aea (teuthology): Fix typo in GitbuilderProject._get_uri_reference()
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 03:01 PM Revision 68b78f7b (teuthology): Add tests for GitbuilderProject._get_uri_reference
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 03:01 PM Revision f7e12c15 (teuthology): Move logic from _get_uri_reference...
- Into new _choose_reference() method. This is because uri_reference is
irrelevant to ShamanProject, but we still want ... - 03:01 PM Revision 82d11183 (teuthology): ShamanProject: split out search URI generation
- Signed-off-by: Zack Cerza <zack@redhat.com>
10/03/2016
- 09:03 PM Revision dfc1e9c9 (teuthology): Merge pull request #960 from ceph/wip-17391
- teuthology-nuke: lazy-unmount osd mount points
Reviewed-by: Dan Mick <dmick@redhat.com> - 04:17 PM Revision fdfaf0d0 (teuthology): nuke: remove ceph data later
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 04:17 PM Revision 04b25e6e (teuthology): nuke: move logging calls, and fix linter warnings
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 04:17 PM Revision 83779307 (teuthology): nuke: Add unlock_firmware_repo()
- Signed-off-by: Zack Cerza <zack@redhat.com>
- 04:17 PM Revision a0c022b8 (teuthology): remove_osd_mounts(): do lazy unmounts
- Signed-off-by: Zack Cerza <zack@redhat.com>
09/30/2016
- 04:09 PM Revision 7d279912 (teuthology): valgrind.supp: tweak
- Signed-off-by: Sage Weil <sage@redhat.com>
- 02:49 PM Bug #17464 (New): AttributeError in fetch_binaries_for_coredumps
- This run is from a little while ago but I think I've seen this more recently too:
http://qa-proxy.ceph.com/teutholog... - 12:25 AM Bug #17391 (In Progress): teuthology-nuke should lazy-unmount fuse mounts
- Here's my PR that hasn't been tested yet:
https://github.com/ceph/teuthology/pull/960
Just need a node or two to ...
09/29/2016
- 09:48 PM Bug #17439 (Duplicate): rerunning of failed tests based on job name
- 09:44 PM Bug #17439 (Duplicate): rerunning of failed tests based on job name
- need ability to rerun of failed tests based on the previous job name which will be very useful
- 03:47 PM Revision ac22f624 (teuthology): Merge pull request #406 from ceph/wip-journal-zero
- ceph.conf: journal_zero_on_create = true
- 03:47 PM Revision 7f463b13 (teuthology): ceph.conf: journal_zero_on_create = true
- Make sure we test this code path.
Signed-off-by: Sage Weil <sage@redhat.com>
09/26/2016
- 10:33 PM Revision 5146ad1c (teuthology): valgrind.supp: another rocksdb leak
- Signed-off-by: Sage Weil <sage@redhat.com>
- 10:31 PM Revision 13dea521 (teuthology): valgrind.supp: a rocksdb Env leak
- Signed-off-by: Sage Weil <sage@redhat.com>
- 03:19 PM Revision e0b7223d (teuthology): Merge pull request #958 from ceph/wip-reweight-config
- ceph.conf: set min rweight min bytes per osd = 10
Reviewed-by: Kefu Chai <kchai@redhat.com> - 03:11 PM Revision 63e2b518 (teuthology): ceph.conf: set min rweight min bytes per osd = 10
- This lets the reweight-by-utilization command work even after mons are
restarted (ceph_manager is currently setting t...
09/23/2016
- 10:43 PM Bug #17391: teuthology-nuke should lazy-unmount fuse mounts
- nuke just succeeded on @smithi042@. So the 'lazy unmount' is all we need here.
- 10:35 PM Bug #17391: teuthology-nuke should lazy-unmount fuse mounts
- ...
- 10:28 PM Bug #17391: teuthology-nuke should lazy-unmount fuse mounts
- ...
- 10:27 PM Bug #17391: teuthology-nuke should lazy-unmount fuse mounts
- Probably unrelated:...
- 10:25 PM Bug #17391: teuthology-nuke should lazy-unmount fuse mounts
- ...
- 10:25 PM Bug #17391: teuthology-nuke should lazy-unmount fuse mounts
- @smithi042@ was last used in http://pulpito.front.sepia.ceph.com/samuelj-2016-09-22_17:41:48-rados-wip-sam-working---...
- 05:27 PM Bug #17391: teuthology-nuke should lazy-unmount fuse mounts
- You left out the real error:...
- 04:20 PM Bug #17391 (Resolved): teuthology-nuke should lazy-unmount fuse mounts
- Maybe new occurrence of #17084
See full log http://paste2.org/Y96czkVh
command line ... - 05:37 PM Bug #16142: Exception during internal.connect fails to unlock machines
- For reference, the sentry event: http://sentry.ceph.com/sepia/teuthology/issues/1341/
- 03:29 PM Revision 6128b7c2 (teuthology): Merge pull request #953 from rjfd/wip-opensuse-pr
- OpenSUSE Leap support in teuthology
- 09:39 AM Revision 8ad5b2b8 (teuthology): install: opensuse: support for zypper package manager
- Signed-off-by: Ricardo Dias <rdias@suse.com>
- 09:39 AM Revision f7959ed3 (teuthology): install: use --capability with zypper {install,remove}
- Otherwise we pick up python-ceph 0.80.9 from the distro repo instead of
a newer python-ceph-compat from the test repo... - 01:22 AM Revision 04a68d85 (teuthology): Merge pull request #949 from dachary/wip-openstack-token-endpoint
- openstack: use OS_AUTH_TYPE=token_endpoint when possible
Reviewed-by: Dan Mick <dmick@redhat.com>
09/21/2016
- 06:04 PM Revision 6fac32f4 (teuthology): Merge pull request #956 from ceph/wip-nodeless
- Set ctx.cluster even when a job has no roles
Reviewed-by: Dan Mick <dmick@redhat.com> - 03:58 PM Revision 45d7222d (teuthology): Set ctx.cluster even when a job has no roles
- This is to allow for teuthology to run self-tests that don't use
resources.
Signed-off-by: Zack Cerza <zack@redhat.com> - 01:31 AM Revision 3e3e62ea (teuthology): Merge pull request #955 from ceph/wip-revert-nodeless
- Revert "Set ctx.cluster even when a job has no roles"
- 01:19 AM Revision 344b7434 (teuthology): Revert "Set ctx.cluster even when a job has no roles"
- This reverts commit ffd14be3bd79baec5d9cafd8972d592085e0e099.
09/20/2016
- 09:28 PM Bug #17351 (New): stale dirs left at the end of the nuke
- /home/ubuntu/cephtest is not cleaned up due to exception collecting coredumps, we should cleanup regardless (-rf) or ...
- 05:38 PM Revision 493af8d4 (teuthology): Merge pull request #954 from ceph/wip-nodeless
- Set ctx.cluster even when a job has no roles
Reviewed-by: Dan Mick <dmick@redhat.com> - 04:12 PM Revision ffd14be3 (teuthology): Set ctx.cluster even when a job has no roles
- This is to allow for teuthology to run self-tests that don't use
resources.
Signed-off-by: Zack Cerza <zack@redhat.com>
09/15/2016
Also available in: Atom