Support passing a template to LXD #505

Closed
castrojo opened this Issue Nov 16, 2016 · 5 comments

Comments

Projects
None yet
4 participants

We'd like to support deploying the canonical-kubernetes bundle on local providers for developers. Currently we need to use another template to get this to work, but Juju doesn't support this:

Idea: could we have conjure just handle the templates automatically so we can deploy kubes to LXD?

@battlemidget battlemidget self-assigned this Nov 16, 2016

@battlemidget battlemidget added this to the 2.1.0 milestone Nov 16, 2016

Contributor

battlemidget commented Nov 16, 2016

This will be addressed sooner than 2.1 just wanted to associate our next milestone with it

Contributor

battlemidget commented Nov 17, 2016

Still working on this, got mostly running except for now getting this error:

Nov 17 02:21:27 juju-74e418-9 kubelet[2361]: I1117 02:21:27.854138    2361 kubelet.go:2237] skipping pod synchronization - [Failed to start ContainerManager [open /proc/sys/kernel/panic: read-only file system, open /proc/sys/kernel/panic_on_oops: read-only file system, open /proc/sys/vm/overcommit_memory: read-only file system]]

You can try this out yourself with

conjure-up -d battlemidget/kubernetes-lxd
Contributor

battlemidget commented Nov 17, 2016

Now fixed with conjure-up/spells#20

sfarosu commented Oct 26, 2017

How did you resolve the "kubelet.go:1757] skipping pod synchronization - [Failed to start ContainerManager open /proc/sys/kernel/panic: read-only file system]" issue?
I'm trying to install openshift origin in multiple lxd containers and i can't get pass this error in kubelet.

I should also mention:
Host machine: Ubuntu 16.04
Lxc containers use centos7 images
Lxd version: 2.19 using snap

My profile is:
config:
boot.autostart: "false"
linux.kernel_modules: openvswitch,nbd,ip_tables,ip6_tables,netlink_diag,nf_nat,overlay
raw.lxc: |
lxc.aa_profile=unconfined
lxc.mount.auto=proc:mixed sys:rw
security.nesting: "true"
security.privileged: "true"
description: Default LXD profile
devices:
eth0:
nictype: bridged
parent: lxdbr0
type: nic
root:
path: /
pool: default
type: disk
name: default
used_by:

  • /1.0/containers/bastion
  • /1.0/containers/master1
  • /1.0/containers/node1
  • /1.0/containers/node2
  • /1.0/containers/node3

schu commented Nov 29, 2017

How did you resolve the "kubelet.go:1757] skipping pod synchronization - [Failed to start ContainerManager open /proc/sys/kernel/panic: read-only file system]" issue?

In case this is still an issue: it should work with ... --config raw.lxc="lxc.mount.auto=proc:rw sys:rw" ... (i.e. change proc:mixed to proc:rw to make sure /proc/syc is not mounted ro separately)

Hope that helps :)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment