Tag Info

New answers tagged

2

Unfortunately, the easiest for now is to access Landscape on the ip address given. You can use something like sshuttle or ssh -D to get access to an environment quite easily with almost no setup. If you need to create an external IP, the IP address you create needs to be accessible to all the nodes, and the outside.Then you can make a change to the ...


0

If you can juggle servers, I would recommend building a new cloud and then figuring out how to migrate the content. A future version of the autopilot (thanks for the feedback :) will support building additional regions in the same cloud, so in future you could make the second region, which would I'm sure be able to migrate glance and cinder content between ...


0

I needed to make certain that port 5240 was available between my desktop and the maas server. It was not open which caused the failure. After making port 5240 available it worked fine.


0

It turns out that it was just a mirror issue. changing from archive.ubuntu.com to one of the mirrors listed here: https://launchpad.net/ubuntu/+archivemirrors ended up working.


0

Thank you for raising this issue. To track status, please check the launchpad bug. Additionally, I am curious if you observe this with Juju 1.24.5? That is the current version available in the Juju stable ppa.


0

After some try and error, I found the only solution to be to uninstall all maas related packages dpkg -r and re-install maas from scratch.


1

When using MAAS and the curtin installer (which is the default since release 1.7) a node install happens in the following set of steps: The node PXE boots into an ephemeral environment using the version of Ubuntu you selected to install. cloud-init downloads user data from MAAS, including ssh keys and curtin. cloud-init sets up the ssh keys and starts ...


-1

This looks like you do not have the directory /var/log/maas just create this directory and re-try installation. it should work.


1

First, for MAAS to be able to manage VMware Workstation VMs, they must exist under the Shared VMs folder in VMware Workstation. Next, rather than adding nodes manually and then trying to set the power type, I would recommend using the Add Hardware > Chassis feature to add the nodes. This is a much more user friendly experience, as it will automatically ...


0

@tai271828's answer does not work for me but gave me an idea instead: To make hung_task became a panic, so kernel can reboot the whole system while encountering a panic. Following are kernel parameters for rebooting if task hung appears: When panic, Kernel will reboot system at 3 seconds kernel.panic=3 Kernel will panic if task hung ...


1

I encountered the same message and issue when I used PXE to install Ubuntu, which happened on precise, quantal...trusty.It could not be reproduced every time, and the reproducing rate is about 1 out of 10 or smaller. So I do not think this is relvant to maas. Could you reproduce this issue every time? This looks like a kernel issue. Please refer to this ...


0

1) It depends on whether you need that cluster to be open to the world. If that is the case, you can either make the switch your "uplink" (plug you r outside network into the switch), or you can set up the MAAS server as a proxy for the other machines. The simplest way to only connect your maas server to the outside world and make it a proxy for the ...


2

Multiple disks are required because the OpenStack configuration deployed by the Autopilot co-locates compute (requiring ephemeral disk) and Ceph (OSD disk for data storage) on the same nodes. You get scale-out storage and scale-out compute in the bargain, but you need two disks. You could substitute USB keys for the second drive... I would not call that ...



Top 50 recent answers are included