Update the zuul status page javascript to use a grey icon when
representing non-live changes. This change has already been made
in the zuul project js.
Change-Id: I60bfa3b7f24e9638fe9be41f07ee472827fb64f9
Apache logs aren't getting indexed after changes to where they exist
in devstack-gate. This means the horizon logs are lost, and that
keystone logs aren't getting indexed at all on the main jobs.
Change-Id: I1ef5084d6bf4dc9f74f4e4b51e00e97573074e38
The zuul user on zuul-dev.o.o could not connect to review-dev.o.o because the
/home/zuul/.ssh/known_hosts file did not get created. This causes zuul-merger
to fail because it needs that to clone repos from review-dev.o.o. This change
passes the gerrit public key from review-dev.o.o so that puppet can create the
known_hosts file.
Change-Id: Iae5d63cc7fd01dc68e4eef96a29bccda6f047eb3
If puppet creates users after any static UID/GID users are created,
it will use the next available UID or GID even if we intend to use
if for another static user. Solve this by increasing the dynamic
range minimums in /etc/login.defs, and also define system account
maximums explicitly so that they don't get interpreted as one less
than the normal user minimums and potentially also conflict.
Change-Id: Ie82ee5cbc23ef38b0eddd8c6af611724cc10fb28
The ansible triggered puppet runs in the puppet_run_all script can force
the script to take longer than 15 minutes if there is much work to do.
This means puppet runs can overlap since the cron executes every 15
minutes. Avoid this overlap if things take a long time with a lock file
enforced by flock.
Change-Id: Ib2680ae3d2641dba7262d09d05a8bcd105b2bb43
This commit removes the 3 unneeded subunit worker processes from the
subunit_worker node type. Previously, 4 workers were launched on the
same node (similar to the logstash workers) but this caused issues
in production. (they consumed all the ram) After dropping down to one
process it was able to handle the load without issue and didn't
consume all the resources.
Change-Id: I5f7f85fc6f8bfb5d2fbec548b58c5d8a1bc0e628
The special py3k-precise nodes are no longer needed by any jobs, so
stop building them and clean up any custom Puppet related to them.
Change-Id: I40d5d09f28ef53583d239d2e852e9c50b5962cf8
Depends-On: Ie105674833e0a527f990a7448855a1090f5dc651
We have a small set of iptables rules on our single use slaves that
enable ironic and heat functionality. We are shifting the floating IP
range from 172.24.4.0/24 to 172.24.5.0/24 and placing an overlapping
range of 172.24.4.0/23 to give compute nodes routes to the floating IPs
in multinode situations.
To accmodate these changes expand the existing rules to cover
172.24.4.0/23 instead of just 172.24.4.0/24.
Change-Id: I0b28c3607747c3939912ce4664627910f431dba6
The spec:
http://specs.openstack.org/openstack-infra/infra-specs/specs/puppet-modules.html
For this particular module, we will take this opportunity to rename the
module to "simpleproxy" to reflect the fact that eventually it will have
no dependency on mysql.
Depends-On: I42597f8bc5ded5b4cdf89b4eae780ae4bcf9f61e
Change-Id: I3cbd5983423ff3a4f59b79af3378c91eff323a49
No longer needed. Remove as part of module-split sprint.
Also remove reference to it from openstack_project/zuul_prod.pp.
Change-Id: I6424d6d7c22b240ca3f1453d5b8770ed56e61e12
The subversion module was a trivial module, so rather than split
it out, just remove the indirection and have the wiki class
install the subversion package.
Change-Id: Ida58a545aba32bb6df65af677f4b4b93c0152851
Rackspace no longer specifies CentOS point releases in their base image
names. This means no more CentOS 6.5 PVHVM image instead we have a
CentOS 6 PVHVM image. Go ahead and use the new non point release
specific image in nodepool to accomodate this change.
Change-Id: I3f9423722615fbacf5ee53a02987fc2052491ef5
Rackspace removed their "normal" non PVHVM base image option. Switch to
using PVHVM to accomodate this change.
Change-Id: I5fc4a40266306e14b733cffa1ed8567369fee7da
In 470f0401 I managed to set the hpcloud-b2 bare-precise build script to
prepare_node_devstack.sh which is the wrong build script for bare nodes.
Change it back to prepare_node_bare.sh so that everything is happy
again.
Change-Id: I96e2bce655dbf036358ee2c485ce549a33fc2b9f
