[04:39] Bug #1578059 changed: Default route not coming up with juju 1.25.5 and bonding

[05:21] Bug #1620838 changed: [2.1a2] Got more than one neighbor [05:30] Bug #1620838 opened: [2.1a2] Got more than one neighbor [05:33] Bug #1620838 changed: [2.1a2] Got more than one neighbor === frankban|afk is now known as frankban [11:27] Bug #1624344 opened: [2.1] Can't SSH to machine in Rescue mode === cueb6526 is now known as chris38 === cueb6526 is now known as chris38 [12:27] Bug #1518414 changed: [1.9] Trusty commissioning (hwe-t) - NVMe model info/serial number is not seen [13:21] roaksoax: Hi. Yesterday MAAS detected one ISCSI node and one VirtIO node. But remaining VirtIO nodes MAAS not detecting https://imagebin.ca/v/2vHCKR5GkVrx. [13:22] roaksoax: I retry the commission again and again. But it was not able to detect disks [13:22] rock_: that seems like a bug then [13:22] rock_: please file a bug [13:25] roaksoax: I deleted all nodes. Again I created 4 nodes with VirtIO disk type. This type MAAS not able to detect disks even for a single nodes also. I recomissioned 5 times. But no use. For ISCSI disk type also the same problem. https://imagebin.ca/v/2vHE0iUlhP07 [13:25] roaksoax: you think , This might be KVM host problem? [13:28] Bug #1624380 opened: [2.1] When passing parameters to maas createadmin, it prompts for ssh-import [13:28] roaksoax: For your setup which company server you have taken? and what was the configuration of that server? Can you please tell me? [13:29] rock_: i don't have a production environemnt [13:29] rock_: i have testing environments [13:29] multiple of them [13:29] rock_: I test 1. kvm's (on a xenial host, VM';s with VirtIO). 2. NUC's 3. servers (a wide variety), IPMI based [13:36] roaksoax: Now I am using KVM VMs on trusty host. This may be an issue? [13:41] rock_: you are using MAAS 1.9 right ? that may be a bug in MAAS [13:42] roaksoax: YES. May be. MAAS 2.0 supports trusty? [13:55] roaksoax: If it supports, For installing MAAS 2.0 packages on trusty what was the best Doc? [14:01] rock_: you cannot install 2.0 in trusty, only xenial+ but 2.0 supports deploying all OS versions [14:01] rock_: and this is likely to be a bug in 1.9 [14:07] Bug #1624403 opened: [2.1] Machines go back to Deployed from Rescue mode, before they actually are [14:11] roaksoax: Can I try like this. By taking KVM host OS as trusty, In one of KVM VM I install Xenial and on top of that I will install MAAS 2.0. [14:25] Bug #1624403 changed: [2.1] Machines go back to Deployed from Rescue mode, before they actually are [14:28] Bug #1624403 opened: [2.1] Machines go back to Deployed from Rescue mode, before they actually are [14:43] Bug #1624424 opened: MAAS not able to detect disks of KVM virtual machines [15:34] greetings all [15:35] gonna try to deploy openstack via maas [15:47] OK, ran into this problem on each of 5 hosts I try to deploy [15:47] each return the cpu count/ram/storage [15:48] and watching the console, each report back with the node name at the login prompt (blade 1 through 5) [15:48] but all report 'failed commissioning' in the web ui [15:49] is there a log some place that I can review the failed installation? [16:14] i have login access to each of the 5 nodes, the local disks are blank, root is mounted via iscsi [16:14] but where are logs stored for the commission step? [16:15] /var/log/maas/rsyslog [16:16] brendand, on the maas server, i assume? [16:16] baldpope, yes [16:18] Sep 16 10:54:02 brk-maas-1 maas.import-images: [WARNING] I/O error while syncing boot images. [16:18] i see that in maas.log [16:19] but the nodes/blades are all up, so not sure if the error is valid? [16:21] would it be worth it to delete and reimport the images? [16:21] rather, sync ? [16:22] baldpope, that means that the images aren't being synced, and I think it's to the region [16:22] baldpope, do you have a proxy? [16:22] kiko, short answer yes [16:23] baldpope, long answer, a complicated one? [16:23] longer answer is that I assume there are two proxies, the internal one specific to maas and then my own? [16:23] correct [16:23] the maas proxy is only used for package installation on the nodes [16:23] ok, then yea [16:23] and only if you haven't specified an external proxy [16:23] if you have specified an external proxy, that's what we tell the nodes to use [16:24] 1sec - let me check network in maas [16:25] ok, i did define the proxy in settings, let me remove that as I don't think I gave network access between the two segments [16:25] okay [16:26] and next, your external proxy may be messing with the large image download [16:26] cool - i'll check 1sec [16:38] i think removing the internal proxy fixed it, updating my notes.. [16:38] thanks kiko [16:38] watching the consoles - it hasn't failed yet -and still shows commissioning in the webui [16:39] would have said failed by now [16:39] yea - blade3 just flipped to ready [16:39] so question - ready doesn't necessarily mean that an OS has been deployed right, if I were to boot, would I see an image on the local disks? [16:40] or is the OS (per blade) always stored on the maas head? [16:43] and (unrelated) ... why can't I paste in the ubuntu session when I run the openstack installer - i keep having to manually type the api key [16:43] frustrating to say the least... === frankban is now known as frankban|afk [16:45] ok, proceeding - bootstrapping juju [16:52] hm [16:53] maas is showing deployed (from deploying) but openstack-installer is still sitting at the bootstrapping juju screen [16:56] baldpope, they are blank still [16:56] baldpope, I know you will find this weird [16:59] so far.. i don't find anything weird [16:59] other than the openstack-install not finishing [17:00] blade1 is up (post deploy) but openstack-install is just sitting here [17:00] from the doc, it should complete and provide a url to start managing landscape [17:01] the only listening port on that is tcp port 22 [17:01] and I can login with the ubuntu user and the public key that was added [17:06] baldpope, but I would not use openstack-install :) [17:06] * baldpope sigh [17:06] baldpope, instead, deploy openstack directly with a bundle [17:06] so... this article - http://maas.io/get-started [17:06] i should just ignore? [17:07] no no [17:07] you did the hard part, which is getting maas set up [17:07] can maas deploy Ubuntu to all the nodes? [17:07] successfully? you can SSH into them, etc? [17:07] that's the last piece of the proof point [17:08] https://insights.ubuntu.com/2015/11/08/deploying-openstack-on-maas-1-9-with-juju/ [17:08] yea - maas so far (getting the blade up and running with no local installation) is gtg [17:08] sorry, i was meant this article -http://www.ubuntu.com/download/cloud [17:09] reading your link now [17:09] baldpope, are you on 1.9 or 2.0? [17:10] anyone knowledgable on maas deployments with lxd? trying to figure out why containers on one machine can't access containers on another. [17:10] l$ juju --version [17:10] 1.25.6-trusty-amd64 [17:10] MAAS Version 1.9.4+bzr4592-0ubuntu1 (trusty1) [17:17] coreycb, yep [17:17] coreycb, what did you do in lxd init? [17:17] baldpope, okay, yeah 1.9 [17:17] kiko, nothing :) [17:17] baldpope, let's do a bundle deploy [17:18] coreycb, heh, okay, lxd init first and choose your networking mode [17:18] kiko, ok - been at this most of the morning - gonna take a quick break for lunch- bb soon [17:19] baldpope, sorry, I have asked that we deprecate openstack-install from the website, because the experience is often sad [17:19] kiko, ok but this is a deploy to maas machines and then containerizing some of the services, so I can't lxd init until after the fact [17:19] coreycb, how is the containerizing being done? by hand? [17:19] kiko, no i'm using the openstack bundle from the charm store [17:20] coreycb, gotcha. and juju 2 or 1.25? [17:20] kiko, juju 2 [17:21] coreycb, can you correctly deploy Ubuntu to the nodes using MAAS alone, and networking between them works? [17:22] kiko, I'd have to check but I'd guess not. they're running on lxdbr0 so I'm wondering if juju should be setting up a different bridge. [17:23] coreycb, so here's the thing: if MAAS can deploy a node with working networking, then any container setup is either the bundle being wrong or a bug in Juju [17:23] coreycb, beta14? [17:25] kiko, ok I'll ask around and see if it's a known issue with juju. it's beta15. [17:58] When I install 1.9.4+bzr4592-0ubuntu1~trusty1 ( of MAAS) it fails. Eearly this week I find out (accidentally ) that 1.9.4+bzr4592-0ubuntu1~14.04.1 ( MAAS) does work. I wanted to force the latter version, but I get unmet dependencies error. How can I solve this? [17:59] This is what I tried: http://askubuntu.com/questions/148638/how-do-i-enable-the-universe-repository#227788 [17:59] This is the error I get: http://askubuntu.com/questions/818368/maas-1-9-4-install-fails-on-ubuntu-14-04-caused-by-a-database-failure [18:00] kiko: maybe u have an idea? [18:01] coreycb: Juju creates bridge interfaces after MAAS has made the deployment, if juju fails creating those interfaces, your machine will be shown as 'deployed' but everything else will be borked [18:02] PCdude: you are not attaching a pastebin on how it is failing :) [18:02] PCdude: one would be ideal to see what's wrong [18:02] roaksoax, did you see the picture? [18:02] PCdude, if you apt-get install postgresql-server before installing maas-region-controller does it work? [18:03] PCdude: that seems like your postgresql is: [18:03] 1. not running [18:03] 2. not accepting connections in localhost [18:04] This is what I got when I tried to install postrgresql-server: http://pastebin.com/raw/cK2dLTCr [18:05] PCdude: sudo apt-get install postgresql [18:06] install postgresql, its is now installing maas again. lets see if it succeeds [18:06] same error [18:07] PCdude: what's it /etc/maas/regiond.conf ? [18:07] PCdude: and what does: sudo service postgresql status tell yo u? [18:08] the first one has exactly one line "maas_url: http://[ip]/MAAS" [18:08] the second one has no output [18:08] just gives me the next line to put my command on [18:10] PCdude: for the message in the image, either postgresq not running or not accepting connections [18:10] PCdude: what does the pg_hba.conf config has ? [18:10] PCdude: make sure postgresql is up and running [18:13] PCdude, I think you can figure out how to get postgresql up and running -- come back to us later when you have figured it out :-) [18:13] yeah, I was trying some other things that could help :) but let me do that first === smgoller- is now known as smgoller [18:26] kiko: roaksoax ok problem solved. BY default version 9.3 is installed. I added the PPA of postgresql and I could update to an higher version. With that version MAAS is now able to install :) [18:27] PCdude, MAAS works with 9.3 though [18:27] PCdude: i think your custom progress doesn't allow connections or doesn't run correctly [18:27] kiko: well not here apparently :) [18:28] PCdude, you are doing something weird [18:28] well all I did was installing ubuntu 14.04 and setup the network after that I installed maas. thats it [18:29] version 9.3 was not running when you tried to install the region controller [18:30] kiko: but that is something that MAAS is responsible for right? I mean, the MAAS installer should install postgresql itself [18:31] PCdude, if you apt-get install maas, you get it installed [18:31] and if you apt-get install maas-region-controller, you get it too [18:32] Package: maas-region-controller [18:32] Depends: maas-region-controller-min (= 1.5+bzr2252-0ubuntu1), postgresql (>= 9.1), rabbitmq-server, debconf (>= 0.5) | debconf-2.0 [18:32] trust me :) [18:33] kiko: I trust you, but apparanetly something went wrong there. I did not installed postgresql before that or had any weird configuration files [18:33] but something triggered it [18:33] PCdude, something is wrong with your system [18:34] kiko: like what? [18:34] kiko, back [18:36] PCdude, no idea. but apt-get install maas-region-controller does set postgresql up for you, automatically [18:37] kiko: and u are sure that maas version 1.9.4 should work with 9.3 of postgresql? [18:38] PCdude, yes, 10000% sure [18:39] kiko: I believe u :) , I have no idea what could have gone wrong. Seriously, its all fresh and no weird things [18:43] PCdude: so you filed this right? https://bugs.launchpad.net/maas/+bug/1618847 ? [18:43] I can close it [18:43] \o/ [18:44] roaksoax: yup, good point. that one can be closed [19:01] Bug #1618847 changed: MAAS install fails, problem with database [19:07] Bug #1618847 opened: MAAS install fails, problem with database [19:10] Bug #1618847 changed: MAAS install fails, problem with database [19:12] uhm mup why are u posting the same error 3 times? :) [19:15] he was filing in triplicate [19:19] kiko, back to that article re: openstack deployment with juju [19:20] baldpope, let's see [19:20] can maas act as a gateway to the intenral cloud, or does the nodes need network access without maas? [19:20] it can act as a gateway, just specify it as the gateway [19:20] and NAT as necessary [19:20] k [19:20] baldpope, and coreycb seems to be following the same instructions [19:20] could be [19:26] kiko, interestingly enough, I found this article while prepping - seems to be a slightly newer version by same author [19:26] baldpope, ah [19:26] grumble. paste not working in hexchat... [19:27] insights.ubuntu.com/2016/01/23/maas-setup-deploying-openstack-on-maas-1-9-with-juju/ [19:29] ah, yes [19:29] this is dimitern's good work [19:41] * baldpope sigh [19:41] kiko, so ... http://pastebin.ubuntu.com/23188223/ [19:44] baldpope, what's the single quotes [19:44] just that - nothing in them [19:44] following the url I posted earlier [19:45] ohg [19:45] i'm retarded... [19:45] nope nevermind.. [19:45] same error [19:46] baldpope, remove the single quotes [19:46] yea - i'm not that bright - i see what I did wrong [19:46] remove single quotes and give it a name before the URI [19:47] yep [19:47] it's been a long week [19:47] baldpope, well, the docs are wrong, right? [19:47] meh - i'm not going to blame the docs [19:48] there are inconsistencies [19:48] i mean step 9 has it written one way, then in his notes immediately below, he has it written an alternate way [19:49] not sure if it's been discussed before, but if it could be noted in the maas client or in documentation that the api key input is masked [19:49] so someone copy/pasting the api key knows that they won't see it pasted [19:49] Bug #1624516 opened: [2.1, FUJ] When removing the ports archive, a weird error message was shown [19:54] the vlan interfaces need to be configured prior to being configured in maas? [20:01] Bug #1624516 changed: [2.1, FUJ] When removing the ports archive, a weird error message was shown [20:03] baldpope, well.. prior to the node being deployed, yes [20:06] kiko, and prior to configuring it maas (which admittedly makes sense, but documentation isn't explicit) [20:07] probably just being picky [20:08] i only make not of it because if you don't do it before installing maas, then it doesn't show up in subnets (the fabrics) [20:10] Bug #1624516 opened: [2.1, FUJ] When removing the ports archive, a weird error message was shown [20:19] Bug #1624522 opened: [2.1, UI, UX] Weird location to put the "Automatically sync images" toggle button [20:46] kiko, fyi I think this is what I'm hitting: https://lists.ubuntu.com/archives/juju/2016-September/007801.html [20:55] coreycb, ah, yes, on 2.0 there is that snag