=== frankban|afk is now known as frankban [09:34] hi everyone [09:34] I am a beginner in maas, and i'm trying to learn and understand the environment [09:35] In the production environment, we have region controller, rack controller, postgresql [09:35] can all of these exist in 1 server? [09:36] and when deploying our first MAAS server, do we need to have 3 physical servers, or shall we use something similar to vm, and create 3 vm servers with these? [09:36] 1 server ==> i mean physical server [09:37] The reason I have the confusion, when we say bare metal, then that means that this is main os that should be installed, and not to run it in a vm, is that correct? [09:39] Also, another question, MAAS does not create VMs [09:46] anyone online? [09:49] Guys? [09:58] is there anyone experienced in maas online? [10:40] do anyone at least know if I could install maas, juju on a single very large (spec) physical machine for production use? [11:17] Bug #1733592 opened: Wrong MTU values for container's NICs [11:26] Dom_, yes, I do exactly that. [11:27] Dom_, in fact, our foundation architecture does that across 3 machines [11:27] Dom_, (it's early for the USA and also it's Thanksgiving week so bear with the delayed responses) [11:27] Hi Kiko, thanks!! [11:27] no, it's alright [11:27] you can use a single physical server for MAAS region and rack [11:28] for Juju, you can bootstrap into a VM on MAAS by using tags or specific placement [11:28] into a VM [on the MAAS node] I meant [11:28] MAAS does create VMs using it's new Pods feature [11:29] Dom_, https://insights.ubuntu.com/2017/10/18/maas-kvm-pods/ [11:29] So I can understand, where does MAAS stand out from esxi for example? [11:29] MAAS itself isn't a hypervisor -- it's a provisioning tool [11:29] Bug #1733592 changed: Wrong MTU values for container's NICs [11:30] I've been into environments where we use esxi to split large physical machines into pieces [11:30] gotcha [11:30] so the KVM pods feature can do something similar [11:30] Hmm [11:30] We are trying to setup a cloud [11:31] as a base, MAAS got our interest, so currently we have 1 large physical machine [11:31] In the future, we are considering expanding the machines [11:31] but is it correct that we deploy Ubuntu MAAS as primary OS? [11:32] correct, Ubuntu and then install MAAS [11:32] And then from there, we advance through the pods feature which uses KVM [11:32] Bug #1733592 opened: Wrong MTU values for container's NICs [11:32] so MAAS technically replaces ESXi with better provisioning, am I correct? [11:34] @Kiki, the reason I'm asking, we're not sure if we should install ESXi as a host, and then create 3 VMs for MAAS, or shall we just go directly for MAAS and let it handle what ESXi does? [11:34] Kiko, the reason I'm asking, we're not sure if we should install ESXi as a host, and then create 3 VMs for MAAS, or shall we just go directly for MAAS and let it handle what ESXi does? [11:38] Kiko, I'm asking from Virtualization perspective? [11:38] Dom_, I'm confused -- can you explain what you are trying to make happen? [11:39] We want to host cloud services & applications, it might be also possible to host full vms, the company has dedicated a single large machine (from spec size) to use a test >> production, if successful, we may get more [11:40] physical machines [11:40] For us to setup this cloud correctly, we noted that MAAS takes care of the physical machine itself [11:40] even though there is confusion about the capabilities [11:41] Dom_, you should get at least 3 machines to do a representative test [11:41] In our environment, we would initially setup ESXi (kind of a very minimized bootable service) then create VM for Windows /or Ubuntu for example [11:42] We are curious if we should really use MAAS directly instead of ESXi >> create VM for MAAS [11:43] Kiko, the test here will be production like environment, then move forward to getting other machines [11:43] Am I still confusing? [11:44] (I'm on the phone so hang in for my response) [11:44] that's ok, tyt [11:44] sorry... === kukacz_ is now known as kukacz [12:35] -.- [14:00] kiko [14:00] I have long phone calls :) [14:02] roaksoax: is there a way to increase the verbosity of the maas logs? [14:11] i'm glad that the batteries lasted that long [14:21] vogelc: rackd.log, maas.log is as verbose as it could be. regiond.log can be more verbose for http requests [14:21] e.g. from the API [14:21] ok [14:22] roaksoax: have you explored using iPXE instead of pxe? [14:38] Bug #1732980 changed: MAAS incorrectly PXE boot UEFI/legacy boot [15:50] Bug #1733900 opened: [2.3final, UI] Machines that have failed testing don't have an error icon <2.3qa> [16:41] vogelc: yes, not a priority atm although we have some implementation proposed there for VM's so far [16:42] vogelc: also, ipxe breaks RFC's [17:46] is it possible to use Powerdns with maas? [17:49] heyya: in what way ? instead of bind ? nope === frankban is now known as frankban|afk [18:09] Bug #1733923 opened: [2.3, HWTv2] badblocks test has 17 badblocks, but test shows as passed. [20:03] Bug #1087183 changed: MaaS cloud-init configuration specifies 'manage_etc_hosts: localhost' [21:00] Bug #1733592 changed: Wrong MTU values for container's NICs [21:00] Bug #1733945 opened: MAAS should warn loudly when it detects mismatched MTU settings [21:09] Bug #1733947 opened: MAAS should allow status transistion from FAILED_DEPLOYMENT to DEPLOYING or COMMISSIONING [21:18] Bug #1733947 changed: MAAS should allow status transistion from FAILED_DEPLOYMENT to DEPLOYING or COMMISSIONING [21:25] roaksoax: I'm running MAAS in vmware. is it possible to use juju charms to deploy windows 10? [21:30] Bug #1733947 opened: MAAS should allow status transistion from FAILED_DEPLOYMENT to DEPLOYING or COMMISSIONING [21:45] Bug #1395904 changed: Need access to curtin install logs from past deployment [21:57] Bug #1395904 opened: Need access to curtin install logs from past deployment [22:03] Bug #1395904 changed: Need access to curtin install logs from past deployment [22:15] Bug #1733947 changed: MAAS should allow status transistion from FAILED_DEPLOYMENT to DEPLOYING or COMMISSIONING [22:23] Hello. I get lshw script failed with status: 139 during every commission [22:23] So they all fail [22:23] Any ideas what causes this?