On 10 May 2017 at 17:46, Leni Kadali Mutungi <lenikmutungi(a)gmail.com&gt; wrote: That one is obsolete AFAIK. The oVirt container images are now being worked on in the 'ovirt-containers' repo. GitHub generally contains a read-only mirror on the oVirt sources, development work for most projects is done via gerrit.ovirt.org. -- Barak Korren RHV DevOps team , RHCE, RHCi Red Hat EMEA redhat.com | TRIED. TESTED. TRUSTED. | redhat.com/trusted

Checked it out. It seems more complicated than what I need, which would be just a Dockerfile that pulls the required image. The ovirt-containers repo seems to have a dependency on OpenCluster, which I think I won't need since I'm after a basic install on one machine for documentation purposes. Your advice is appreciated. -- - Warm regards Leni Kadali Mutungi

On 05/11/2017 08:54 AM, Barak Korren wrote: You may be able to use the engine container independently, with some effort. But this 'ovirt-containers' project isn't certainly intended for that simple use case, it is intended to be able to deploy the complete oVirt system in an OpenShift cluster. If you are willing to give this a try, and you already have a docker installation, creating an OpenShift cluster for testing purposes is really simple. You can use 'oc cluster up'. It is a matter of downloading the 'oc' tool from here: https://github.com/openshift/origin/releases/download/v1.5.0/openshift-or... That contains the 'oc' binary. Put it in your path and then run this command: oc cluster up That will pull all the required images, and configure a simple OpenShift cluster into your machine. Once that completes you can proceed to deploy the oVirt application. First you need to build the images, as they aren't yet available in a public repository. To build the images checkout the code of the project: git clone git://gerrit.ovirt.org/ovirt-containers Then go into that directory and run this: make build That will take quite a long time, as it needs to build all the images, and it requires downloading of the base images and of the RPMs that go into the images being built. Be patient. Once the images are built, you can deploy the complete oVirt application running this: make deploy Would be great to have you testing this, giving feedback and reporting any issues you find. But be aware that this is in a very early stage of development and should be considered experimental.

On 05/15/2017 05:24 AM, Leni Kadali Mutungi wrote: The tool that builds and deploy the containers to the OpenShift cluster is a program written in Go. It is compiled when you run 'make ...'. That means that you need to have the go development environment installed. For example, in CentOS or Fedora: yum -y install golang Once you have it installed, try again 'make build' and 'make deploy'. I'll try to update the documentation to explain that.

On 05/18/2017 05:29 AM, Leni Kadali Mutungi wrote: The problem is that the Makefile uses the current working directory to setup the GOPATH. When you use 'sudo' the current working directory is the home of the root user, hence the Makefile won't work. Please try to run the build with a non-root user, just go into the the source directory and type 'make build'. You will also need to make sure that the user has docker permissions, if you didn't do it yet. To do so just add it to the 'docker' group. For example, in Fedora or RHEL: usermod -aG docker youruser Note that once you do that you will need to re-login to make it effective.

On 05/19/2017 05:20 AM, Leni Kadali Mutungi wrote: Did that restart of the docker daemon resolve the problem? As far as I can tell that 'RUN sed' instruction that failed is correct. I just re-checked it in my environment and it works correctly. Looks more like a resources issue or a bug in docker itself. The message says OOM, out of memory, is that machine low on memory maybe?

On 5/22/17, Leni Kadali Mutungi <lenikmutungi(a)gmail.com&gt; wrote: On 5/19/17, Juan Hernández <jhernand(a)redhat.com&gt; wrote: Seems to be a problem related to the Debian Kernel. cgroup swap accounting is not enabled. I tried adding `cgroup_enable=memory swapaccount=1` to the GRUB_CMDLINE_LINUX_DEFAULT option in /etc/default/grub after the `quiet` option. After this I ran `sudo update-grub` and restarted my computer. The memory error has disappeared, the RUN sed instruction error has not. Step 26/34 : RUN sed -i 's/multipathd.service //g' /usr/lib/systemd/system/vdsmd.service ---> Running in 67d580645333 sed: can't read /usr/lib/systemd/system/vdsmd.service: No such file or directory The command '/bin/sh -c sed -i 's/multipathd.service //g' /usr/lib/systemd/system/vdsmd.service' returned a non-zero code: 2 Failed to build image 'ovirt/vdsc:master' Is vdsm a dependency for ovirt-containers? If so, how do I get the container to successfully locate the service it seems to be looking for? CCing Oved Ourfali since Juan Hernández is on vacation till May 29th. -- - Warm regards Leni Kadali Mutungi

This is awesome, I've been playing w/ this on a centos host running origin 1.5.1. I've gotten stopped when I've tried to approve the host in the engine, that fails with an error message like: "Host vdsc-ds-615nl installation failed. Failed to get the session.." Any idea what what's going wrong there? Jason