Hi, You might be aware that OpenShift team is restructuring/re-architecting the "oc cluster up" code base in v3.10. Hence in Minishift we are trying to adapt these changes [2]. In the recent discussion with OpenShift developers we came to know that the remote docker support in "cluster up" is fragile and it will bring more trouble for future OpenShift features which needs to be enabled in "cluster up". Refer [1] for more details. Also one of the major reason the remote docker daemon support is present in "cluster up" to support Minishift. Minishift uses the remote docker scenario i.e. the docker daemon runs inside the Minishift VM and we use the oc binary on the host to run the "oc cluster up" using the remote docker daemon in the VM. One of the important expectation of Minishif users is around ability to use/try new features of OpenShift in Minishift. It seems the remote docker daemon support in "cluster up" is going to make things difficult or hard to implement new features to "cluster up". Which will also make things more prone to bugs, bad user experience etc. I would like to suggest that we drop the support for remote docker in Minishift. However doing so will make the code backward incompatible to earlier versions of OpenShift and couple of features (e.g. add-ons) needs to be checked and modified to suit to the new implementation. This will also make things more simple/maintanable for OpenShift developers too. So we should go ahead with Minishift 2.0.0 (major version change) and if required we can maintain the previous version of Minishift i.e. v1.x for sometime (only blocker bugs or security fixes) as will help if someone needs to run older versions of OpenShift with Minishift. Sugestions, comments? [1] https://github.com/minishift/minishift/issues/2377 [2] https://github.com/minishift/minishift/pull/2336

Lala

Hi, You might be aware that OpenShift team is restructuring/re-architecting the "oc cluster up" code base in v3.10. Hence in Minishift we are trying to adapt these changes [2]. In the recent discussion with OpenShift developers we came to know that the remote docker support in "cluster up" is fragile and it will bring more trouble for future OpenShift features which needs to be enabled in "cluster up". Refer [1] for more details. Also one of the major reason the remote docker daemon support is present in "cluster up" to support Minishift. Minishift uses the remote docker scenario i.e. the docker daemon runs inside the Minishift VM and we use the oc binary on the host to run the "oc cluster up" using the remote docker daemon in the VM. One of the important expectation of Minishif users is around ability to use/try new features of OpenShift in Minishift. It seems the remote docker daemon support in "cluster up" is going to make things difficult or hard to implement new features to "cluster up". Which will also make things more prone to bugs, bad user experience etc. I would like to suggest that we drop the support for remote docker in Minishift. However doing so will make the code backward incompatible to earlier versions of OpenShift and couple of features (e.g. add-ons) needs to be checked and modified to suit to the new implementation. This will also make things more simple/maintanable for OpenShift developers too. So we should go ahead with Minishift 2.0.0 (major version change) and if required we can maintain the previous version of Minishift i.e. v1.x for sometime (only blocker bugs or security fixes) as will help if someone needs to run older versions of OpenShift with Minishift.

Sugestions, comments? [1] https://github.com/minishift/minishift/issues/2377 [2] https://github.com/minishift/minishift/pull/2336 Thanks, Lala _______________________________________________ Minishift Mailing List mailing list -- minishift(a)lists.minishift.io To unsubscribe send an email to minishift-leave(a)lists.minishift.io