Minor tweaks also include:
- shoebox handler now wraps notification with region/cell info
- pip_freeze_versions.txt is updated after each ./build and
copied into the VENV directory for inclusion in tarball.
Change-Id: Ibe4027770b859ea7878c9f0952fccf94fb496c43
Also added support for quick build (does not reinstall everything in .venv)
which is useful when you have to hand-tweak a venv for testing. Generally
though, it shouldn't be needed.
Change-Id: I4b93c8e256f49d16d2b0fa821b37355526885ce8
* Update Vagrantfile to run git clone command as unprivileged to avoid having root own the directory
* Update README to reflect changed command
Change-Id: Ibfde4f9cce2367ee44afb43dbafbaf6d626d7381
For uploading raw notifications to archive (like an HDFS store).
Also, removes msgpack-python (installed by oslo), which currently
conflict with carrot. This is a temp fix.
Change-Id: I973949c83c9018235169646ce0370d71a2260241
* Add simple Vagrantfile to provision an Ubuntu 14.04 VM with the bootstrap.sh script
* Update README with instructions for use
Change-Id: I94cd69742aeffab4c01a27510c25f381151bfefe
Also backports some tweaks from ops:
- fixes logging.conf
- fixes pipeline_worker first time start
Change-Id: I6ab1e567594097bc23d857c05d0c1fc84262babf
In order to run StackTach.v2 and StackTach.v2 in parallel,
we need to have two queue from OpenStack (vs. having to change
--notification_topics to include a new queue).
This approach makes a new queue with the same routing_key off the
same exchange, resulting in mirrored queues which can be consumed
separately.
Exchange "nova" -> (routing key "monitor.info") -> Queue "monitor.info"
-> (routing key "monitor.info") -> Queue "monitor.info3"
Change-Id: Icce87c7eb43025672759d442553feb325e186128
Looks like we were using the default config files in the ansible
scripts when we should be using the Usage related ones.
Change-Id: Id30152ab7762aff85f3258552de0b5a3e0a6e876
Utility for extracting worker and rabbit information from
large STv3 deployments. Easier than having to check many
different hosts. See etc/status.yaml for sample config file.
Change-Id: If1b39ce3b34235f3bf0eac4ea707be2087ca14ad
Yagi-workers are now created per cell. One rabbit per cell.
Also renamed the common role to stv3-common to prevent conflicts
with other playbooks.
Take note of ansible/roles/stv3-workers/defaults/main.yaml to
see an example of the expected format of the calling facts file.
Each entry in this list-of-dictionaries represents a rabbit server
and Swift container for that cell.
Also, multiple pipeline_workers are created based on the value
of num_pipeline_workers (defaults to 2). This is so a machine
can better utilize multi-core architectures.
Change-Id: I09c4d364878493546d6e23e019f01f5f4f38c37c
Services now run properly on first run (thanks Joel). Didn't
need to check the changed status of the command after all.
Also, spelling mistake in the playbook launch fixed.
Change-Id: I5b6cb276b066844c934da4ffbd2513a006033214
Seems to be a change in how infra defines the stackforge repo.
Also removed a stv3-api restart from common, which was undefined.
Change-Id: I3c33dd6db467225444a0a9da137bf8f002629e3a
Optionally run tox on each git repo (-t) and package resultant
venv into a tarball on success (-p)
tarball filename includes the short-sha1 of the repo.
If tarball is created, screen is not run.
if the -d option is given the cmdline incantation for kicking
off the ansible deploy playbook will be provided. Later, this
could be automated.
A mysql role playbook is also included for single node db.
stv3-api role is also included.
Change-Id: Ib9f97a0035c228d4a163d7794ab7ce60043d572b
