We do run Karaf inside docker (using static assemblies). The best thing I see in it is possibility to SSH into karaf and see what's going on. This is especially useful when logs are not dumped to stdout nor file but sent via network somewhere else. Log tail then shows all events before they are serialized.
A nice side effect of SSH console is possibility to execute commands via exec. ;-)) By this way you can browse basic file system contents and verify files delivered via container/cloud orchestration tools. Having said that we run static assemblies, I would like to note that we do not rely on "dynamic" updates of configurations just yet. We experiment with kube as new way to run things but we didn't reach yet place where we deliver configurations from etcd. This is place which will get really interesting. Most likely it will require us to go back to traditional, non static, assembly in order to pick up updates pushed via kube api. We do have separate pipeline (with simple makefiles) for building docker images which is detached from main build where we produce custom distro. Its used to add cloud specific configurations. In two projects where we didn't use a common configurations we had extra profile in maven build for that. Developer can run static assembly locally with minor inconvenience caused by static configs. For local builds we also used docker-compose which allowed to setup all required dependencies (such database). Cheers, Lukasz On 8/18/18 11:00 PM, Christian Schneider wrote: > I am currently looking into optimized ways to run OSGi applications in > the cloud. > I would like to describe how cloud native OSGi could look like. > > Apart from my own experiments I would like to hear from other karaf > users. Do you run OSGi and especially karaf applications in the cloud? > > How do you build your application? How do you do releases and > deployments? What do you do different compared to a non cloud setup. > > Regards > Christian > > -- > -- > Christian Schneider > http://www.liquid-reality.de > > Computer Scientist > http://www.adobe.com >
