Turns out we got the wrong hash before :( but thankfully reprepro didn't
funciton in that state. Get the ids correct so reprepro can do its job.
Change-Id: Ic8e5f3ebfea681b289e1f5381134df4e0485af3d
Mirror the trusty packages for the ceph hammer release to aid in process
of making ceph testing more robust. Use reprepro which is already in use
to mirror the main ubuntu trusty repos.
Change-Id: Ifd09272c7b1e07de9135be5a96be06153a3f611e
The puppet-subunit2sql module updated where it puts config file to
be in /etc/subunit2sql instead of /etc/logstash. But we missed updating
the worker yaml config file. This patch corrects the oversight.
Change-Id: Ibb4d2eee7deb7876fc17c49dade8d4ae1236ca3e
Due to various puppet changes which were required to support running
OpenStackId on Trusty, I7f1434a6218d3300d1315a2e3c0a446f05124353
disabled Puppet updates on the production openstackid.org server
temporarily. Given that the current plan is to replace
openstackid.org with a new server running Trusty now, switch the
disablement from its name to its UUID so we can successfully puppet
the replacement server.
Change-Id: I37fe3763c1e27ff7dfea2e2012f5df071224d30a
A refactor for the puppet-openstackid module in
I66c6ad413a6b0c31a19cc663058a53edc3bec5cc (switch to Puppetlabs
apache module, apache2-mpm-event, php5-fpm proxy handler) for
improved performance needs to be vetted on
openstackid-dev.openstack.org without adversely impacting production
openstackid.org.
This temporary disablement will be reverted once the updates are
confirmed to have the intended effect on the dev server and not
subtly break it in ways manual testing and CI jobs have not exposed.
Change-Id: I7f1434a6218d3300d1315a2e3c0a446f05124353
Stop running Puppet on openstackid.org so that we can manually apply
a hostname fix in concert with the puppet-openstackid change
I41ddb3cd12fa564d19cedc18ca5585d2ca7481da .
Change-Id: If67783a2cbaf77e89a3d2aae80a9206b547dbac6
Ubuntu isn't the only thing we want to mirror, we may also want to
mirror debian or ceph packages and so on. Refactor the puppet so that it
is easier to do this in a reconsumable way.
Change-Id: I0a12bc4cb67339a7566fb113bbbc897d4f112f50
According to [1], UIDs and GIDs between 0 and 200 are reserved for some
services, including OpenStack services.
Puppet OpenStack is deploying OpenStack using RDO packaging and there is
currently a race condition where sometimes RabbitMQ is installed
*before* Keystone and steals its UID, Keystone fails to be installed,
and CI job is failing.
This patch prevents this situation by setting both SYS_UID_MIN and
SYS_GID_MIN ti 201, so packages using dynamic UID/GID won't steal
reserved ones.
[1] https://git.fedorahosted.org/cgit/setup.git/tree/uidgid
Change-Id: I7cc41adc7e6070c32d9a56940a67640ab4419e7e
As they come from an static inventory, we need to
add all the servers into the list. At the moment, adding
them as disabled, because they will need that all pending
changes are merged until puppet can safely run on it.
Change-Id: I665b12912d38562230e2bb8d53a35dbc37dd9c06
This commit adds back the job filter on the old side subunit stream
from grenade runs. The subunit stream will only end up under logs/old
if it's a grenade run. So to be more efficient lets filter looking
for a subunit stream there only to grenade jobs.
Change-Id: I5a6c9b17f923526745507a131658f581a9d94e0a
This commit removes the retry-get parameters from the subunit file
collection in the logstash gearman client. These were cargo-culted
from elsewhere in the yaml file and doesn't add anything useful in
this case. The param is only there to workaround jenkins limitations
with the console log uploading.
Change-Id: Ic751d5e295e605b20386ad9cde6500bae44f34d8
This commit will enable collection of subunit results from all subunit
emitting jobs in the gate and periodic pipelines. Now that we prune
the database to only 6 months of history we should have plenty of
headroom for the extra data generated by non-tempest-dsvm jobs.
(hopefully)
Depends-On: I58a640f804313e1e4b80680f0e39b86d76cb29da
Change-Id: If5691af792409f02352f25b1498dd78294a7cd74
Some infracloud machines do not respond to ssh. Comment them out here
so they don't fill up our logs with errors that aren't actionable.
Change-Id: If56e773e3fe8beb0aad42e079eb8cb8492cde652
Add separate playbook for infacloud nodes to ensure they run in the
correct order - baremetal -> controller -> compute.
Baremetal is intentionally left out, it is not ready yet.
All 'disabled' flags on infracloud hosts are turned off. This patch
landing turns on management of the infracloud.
Co-Authored-By: Yolanda Robla <info@ysoft.biz>
Co-Authored-By: Spencer Krum <nibz@spencerkrum.com>
Change-Id: Ieeda072d45f7454d6412295c2c6a0cf7ce61d952
We only need one infracloud right now. Technically the controller
doesn't need to be in 'infracloud' because its secrets are duplicated.
This aligns the groups at any rate.
Change-Id: Ibc056505a4ef397c2adeae6a3b4afb21db9b4a02
k5start prints:
"Kerberos initialization for service/reprepro@OPENSTACK.ORG"
This then gets redirected into the list of files to remove on the next
run - which is an error, since that's not a filename. Put the
redirection inside of a bash command so that only the data gets
redirected.
Change-Id: Icae10763b4bce78c7b674715826547c228980dd6
We can't colocation debian and ubuntu - so allow for each having their
own pool independent pool structure.
Change-Id: I9fe0fa9769a5c0f6a852844b2db0dfb620e91375
The packages in restricted and multiverse are either non-free or
unsupported. While they're usually useful for desktop, they aren't
something that OpenStack should grow a dependency on.
Also remove non-free and contrib from debian. We don't do anything with
debian yet - but if we did, I'm pretty sure we would not want these.
Note: The existing restricted and multiverse dirs should be purged,
although it's not urgent to do so. 'reprepro --delete clearvanished'
Change-Id: I1d98ab1dea52598d01e3ce3d9a24306280a190fe
Now that the db files are in AFS, we need k5start because we need to be
able to write lockfiles for them.
Change-Id: I5c9554ca505ed56e84b601e40476527ef4af98fb
We need the gpg signing keys for the upstream repos in root's keyring.
Fuel uses golja-gnupg for importing keys - so seems like a good thing
for us to use too.
Change-Id: I343be9397ba53f494485d1ff0f1e1e4c96c0273f
In the current setup, the reprepro db and lists dirs are being written
into /root. Move them into AFS so that we store all the data in the same
place.
Also, update the script to point to the confdir for reprepro commands.
Note: This will require a manual move of the current db and lists dirs
from /root to /afs/.openstack.org/mirror/apt before this is run.
Change-Id: I2638621e6af98296704e680227fae174d8a0e541
Since these are baremetal hosts, they need to come from a static
inventory not the openstack inventory. Fortunately, thats pretty easy.
Also setting infracloud groups to be children of disabled to keep them
disabled until we are ready.
Change-Id: I87ed4008ed9c4867f79bbb5fbb6be53707b42625