github
/
k3s
mirror of https://github.com/k3s-io/k3s
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Document package default for etcd persistence

pull/6/head
Dr. Stefan Schimanski 2015-12-08 16:32:23 +01:00
parent 41cb3d6a39
commit 47a541aa1e
1 changed files with 10 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
contrib/mesos/docs/issues.md
View File

@ -198,6 +198,16 @@ The scheduler also offers `/debug` API endpoints that may be useful:
All of the issues in the above section also apply to the Kubernetes-Mesos DCOS package builds.
The issues listed in this section apply specifically to the Kubernetes-Mesos DCOS package available from https://github.com/mesosphere/multiverse.
### Etcd
The default configuration of the DCOS Kubernetes package launches an internal etcd process **which only persists the cluster state in the sandbox of the current container instance**. While this is simpler for the first steps with Kubernetes-Mesos, it means that any cluster state is lost when the Kubernetes-Mesos Docker container is restarted.
Hence, for any kind of production-like deployment it is highly recommended to install the etcd DCOS package alongside Kubernetes-Mesos and
configure the later to use the etcd cluster. Further instructions
can be found at https://docs.mesosphere.com/services/kubernetes/#install.
This situation will eventually go away as soon as DCOS supports package dependencies and/or interactive package configuration.
### Kubectl
The following `kubectl` and `dcos kubectl` commands are not yet supported: