Speedup Kolla with Ansible fact caching

By | 10.08.2016


The Ansible master playbook of the Kolla project is big and consists of a lot of playbooks. The fact gathering before each call of a playbook requires time. The required time depends on the size of the environment and is possibly not negligible.

Fact Caching, introduced in Ansible 1.8, helps avoid the side effects of excessive coffee consumption while waiting for the execution of actions. Ansible ships with two cache plugins: JSON and Redis. The following example highlights the Redis configuration.

Configuring Fact Caching

To set up Redis fact caching, proceed as follows on the Kolla deployment host.

  1. Start a Redis container:
  2. Install the Python client for Redis:
  3. Add the following code to the Ansible configuration file, e.g. /etc/ansible/ansible.cfg:

Testing Fact Caching

Use the following playbook for a quick test:

Save the playbook to testing.yml and run it with the inventory file of the Kolla environment. The execution time should be fairly short, once the facts have been initially cached prior to the first run.

Vote for Presentations for Barcelona OpenStack Summit

By | 01.08.2016

Sagrada Familia in Barcelona

Sagrada Familia (Bernard Gagnon, CC BY-SA 3.0)

Vote for Presentations for the Barcelona OpenStack Summit in October is now open. Pick your favorites from more than 1500 proposed sessions!

Unfortunately, the voting interface does not support to link to presentation proposals this time. If you intend to vote for a specific presentation, you have to manually search for either speaker or topic.

Below is a list of our submissions for the Barcelona OpenStack Summit.

Non-public public cloud with confidence

Speakers: Christian Berendt (Betacloud Solutions), Kurt Garloff (T-Systems)

Companies increasingly look at public clouds also to process highly confidential data. To facilitate this, public cloud providers need to invest into solid security protection.

Security specifications in an isolated private environment are a challenge. For a public accessible environment they are a nightmare. In this session we will address security topics from the perspective of public cloud providers.

After starting with common & best practices, like the importance of a central logging with a good set of queries and the use of VPNs to access internal networks & applications without exposure to public networks, we will have a look on security tools and frameworks, inside and outside the OpenStack ecosystem.

Documentation, business processes and education are other important topics that need to be revisited in the context of security.

The presentation will give a look behind the scenes of Open Telekom Cloud (an OpenStack based public cloud) with respect to those security aspects.

My spider sensu is tingling!

Speakers: Christian Berendt (Betacloud Solutions), Jan Klare (cloudbau)

Most of the OpenStack core projects are currently considered stable and can be installed plug-and-play fashion. Although most of the time everything seems to work fine, anyone operating an OpenStack environment dreads errors appearing silently in the background where they might stay undiscovered while they are not affecting anybody. Some of these might not be worth looking into, but some of them might cause serious trouble later on.

During the last few years, many monitoring solutions have evolved around OpenStack or plug-ins for existing ones have been developed. This session introduces a few Open Source options for monitoring your OpenStack cluster. We will outline the monitoring solution we are currently using at cloudbau and Betacloud, which is based on Sensu, InfluxDB, Grafana, Elasticsearch and Collectd. We will finally conclude why we think that Open Source and community-driven solutions are best suited for monitoring OpenStack environments.

How to charge your clients?

Speaker: Christian Berendt (Betacloud Solutions)

After the deployment and the integration of an OpenStack environment it is a common use case to charge the users for provided and consumed resources, regardless of whether it is a public or a private cloud. While there are feasible metering services in OpenStack, billing and rating based on the metering data still is a major nuisance. We will outline the existing metering and billing solution we evaluated for the use at Betacloud and show how we integrated one solution with an online service for quoting and billing.

5 Years of OpenStack Managed Service at SAP SE

Speakers: Christian Wolter (B1 Systems), Boris Wuest (SAP)

In 2011, B1 Systems began evaluating OpenStack vs. SAP’s traditional script-based Xen hosting solution. The projectstarted off as a PoC based on OpenStack Bexar. In 2012, B1 built a fully-fledged OpenStack production environment with +200 hypervisors and 135TB of memory.

For more than four years, B1 has been operating and providing support for SAP’s OpenStack landscape.Since then, a total redesign of SAP’s OpenStack landscape has taken place in order to benefit from OpenStack’s fast development pace. The migration of two thousand VMs from Folsom to Icehouse was but one of the major milestones. This environment features OpenStack customizations, a workflow engine and an enduser frontend (B1 Cloud Portal) tailor-made for SAP, thus integrating the new environment with thousands of VMs seamlessly intothe existing SAP infrastructure.

SAP and B1 Systems will share their OpenStack journey – including their experiences regarding customizations, operation and upgrades.

Images auf OpenStack zwischen Projekten teilen

By | 21.07.2016

In vielen Situationen ist es sinnvoll Images nur innerhalb bestimmter Projekte bzw. einzelnen Tenants bereitzustellen. In OpenStack gibt es dafür ab der Image API 2.1 die Members Funktion. Der Workflow ist im Detail im Image Guide der OpenStack Dokumentation beschrieben.

Image teilen

Um ein Image nun mit einem anderen Projekt zu teilen benötigt man zunächst die ID des zu teilenden Images, welche mit dem Kommando openstack image list abgerufen werden kann. Weiterhin wird die ID des Zielprojekts benötigt (openstack project list).

Anschliessend kann mit dem Kommando glance member-create das Image dem Zielprojekt zugewiesen werden. Aktuell ist es noch nicht möglich diese Funktion mit dem python-openstackclient zu nutzen.

Geteiltes Image akzeptieren oder zurückweisen

Der Status ist zunächst pending. Im Zielprojekt hat man nun die Möglichkeit das zur Verfügung gestellte Image zu akzeptieren, über den Status accepted, oder zurückzuweisen, über den Status rejected.

Mitglieder eines geteilten Images anzeigen

Mit dem Kommando glance member-list lässt sich der Status aller Members eines Images anzeigen.