[02:36] hi so I am doing my first PoC deployment and I am trying to understand why I am having problems getting juju to deploy lxc containers [16:45] Hi, I am trying to use charm actions, and all the charm actions seem to be failing with “Permission denied”. [16:45] root@juju:~# juju action fetch 7b9b575d-7853-49d9-82ae-7b9fd3d1a6f7 [16:46] message: 'exec: "/var/lib/juju/agents/unit-cwaio-cwims-vnfd-b-3/charm/actions/create-update-user": [16:46] permission denied' [16:46] status: failed [16:46] timing: [16:46] completed: 2016-02-14 16:35:42 +0000 UTC [16:46] enqueued: 2016-02-14 16:35:41 +0000 UTC [16:46] started: 2016-02-14 16:35:42 +0000 UTC [16:46] Appreciate any help on this. [18:45] This may be more of a maas question, but it's also a bootstrapping question. I have recently installed 40Gbe infiniband with switch {Mellonox} between a homebuilt server/ ZFS NAS, which is hosing my MAAS nodes. Nodes are connected to the internet via stand alone PFSense box, one netgear 24port prosafe POE, and a Cisco 48 port managed switch. PFSense is forwarding all BootP requests to MAAS box, [18:45] but handling all other DHCP/DNS requests. IMPI works for enlisting and initial provisioning of nodes. I am wanting to put the Jujubox into a VM, pass both the Gbe Lan and IB networks to the VM, then be able to deploy clusters, or openstack via JUJU using the IB network for the storage network, lan for command and control/inet I have an R900 with 4 hex core 1.8Ghz processors, 120 Gb ddr2 5300F, [18:45] connected to a Tesla S1070, and Tesla S2090. linked to IB and Ethernet, 2 Dell C6105 3 node servers, 2 Dell 2950s, 2 reporposed dualcore nodes, and a couple of home built boxes. All set to boot from pxe over the lan link, but also connected to the IB network. MAAS is handling DNS/DHCP for IB subnet, but not eth0. I have gotten a VM enlisted and configued in MAAS but it times out when trying to [18:45] bootstrap Juju, I have already set the juju timeout to 3600 in the config. Thoughts? [19:05] anyone? [20:01] can any help understand what ERROR saving bootstrap endpoint address actually means? [21:29] is there a way to get juju to “redeploy” a machine and it’s services? [21:34] stormmore: I generally rely on `juju destroy-machine --force, juju upgrade-charm, juju add-unit [21:35] would that work for bundles too? [21:35] stormmore: there may be a more elegant way to achieve the same thing, I'm uncertain. [21:35] that is still more elegant that destroy environment and rebuild when one machine doesn’t come up in time [21:37] right, much faster than that. [21:37] I have already lost a machine this weekend as it won’t respond to wakeonlan anymore [21:40] I blew the lab manager’s mind when I showed him MAAS last week [21:40] hah yeah, MASS is pretty cool. [21:40] err MAAS even. [21:41] it was what he wanted a few years back and never got [21:42] but wow does juju really make maas powerful :) [21:42] my mine was almost blown away with juju-gui [21:43] and that is is hard to say about any modern tech for me [22:09] love watching the juju status --format tabular when launching a bundle using quickstart === blahdeblah_ is now known as blahdeblah [23:22] now if only I could get the networking right on this proof of concept [23:23] don’t think I have enough routebale IPs for it at the moment... only have 12 [23:57] stormmore: yeah --format=tabular is so awesome it'll be the new default in 2.0 [23:59] agunturu: are the files in actions/ executable (chmod +x)?