[00:17] hey someone can help me in something [00:17] im setting up a computer over kvm [00:17] with maas [00:17] one of the interface should use a vlan [00:17] but is not taking the vlan setting on it [00:17] any idea of why? [00:19] but in baremetal hardware create the interface eno1.vlanid [00:29] jose-phillips: do you have an vlan created in the networking model in MAAS under the same fabric where you need to create it ? [01:27] roaksoax: any other ports might need forward to the region from the proxy? [02:02] Bug #1716750 changed: VM deployed by MAAS has limited network functionality [02:36] guessing I need the 5248 forwarded too [02:44] roaksoax yes [02:45] roaksoax on the baremetal server wokrs as expected [02:45] on virtual machine i select just 1 interface with vlan but when the computers get deployed [02:50] jose-phi_: what vm, a pod based VM ? [02:50] kvm pods don't support fancy networking [02:50] is nto a pod [02:51] is a virtual machine and looks like a physical computer [02:52] i see the deployed vm and check cloud-init config file inside of /etc/networkings nothings point to a vlan [02:54] jose-phi_: that would only mean the machine was not configured correctly in MAAS [02:55] jose-phi_: maas passes the configuration you make in the MAAS ui [02:55] to deploy the machine [02:55] maas uses the machine as any physical system [02:55] so you'd have to configure it in MAAS [02:57] roaksoax https://pasteboard.co/GVFRKmS.png [02:59] https://pastebin.com/5nUSQjGd [02:59] there is the config fil [02:59] e [03:00] on networking [03:00] jose-phi_: i dont see any interface configured to enoX.vlan [03:00] exactly [03:00] that is a machien deployed with maas [03:00] machine [03:01] with the screenshot i send you before [03:01] jose-phi_: the config that you showed me makes sense to me [03:02] jose-phi_: you need to create a vlan interface [03:02] jose-phi_: e.g. go to the interface , on the 'Actions', select the 3 vertical bars [03:02] but maas don't should create it? why works with the physical server [03:02] jose-phi_: 'Add alias or VLAN' [03:03] select 'VLAN' [03:03] and create a vlan interface [03:03] jose-phi_: in your setup MAAS has not idea that you are running KVM on a maas deployed machine [03:03] so the VM's wont inherit the host networking [03:04] ahhh got it [03:05] so i can leave [03:05] the physical interface [03:06] unassigned [03:06] and create the vlan interface over it [03:06] right? [03:28] https://bugs.launchpad.net/maas/+bug/1734798 [03:28] thanks you so much roaksoax i will try this setting [03:38] Bug #1734798 opened: machines never deploy [04:32] Hello, I was hoping someone could tell me if MAAS is compatible with iLO3 or if there was a way to set it up? I'm attempting an OpenStack installation on HP DL380 G7 servers, and having to use manual power type. Thanks in advance! [04:57] With using Manual power type, I've been able to successfully run a deploy. 3 servers get the Ubuntu image but when I get to installing the remaining OpenStack applications it hangs on "waiting for deployment to settle". Its been sitting for about 30min now, nothings changed. Think I'll let it sit over night, thanks again! === frankban|afk is now known as frankban [14:23] Hi all, I have some problems with a cluster setup here. Anyone willing to take a look? [14:24] (Do note that I am a beginner in this topic) [14:26] https://askubuntu.com/questions/981114/maas-with-external-dhcp-or-static-assignment [14:44] how's it going, all? [14:44] I have an issue where my machines never deploy [14:44] I've created a bug for it [14:45] wondering if there might be anything obvious someone might want to point out [14:45] I took the proxy out of the picture [14:45] I just have a single region+rack [14:45] just trying to get things to work in singleton mode first [15:00] bdx: never deploy means it finishes image instal, but when it reboots, it does not boot onto the OS ? [15:00] bdx: or it fails the image install? [15:35] roaksoax: it boots to the os and just sits there [15:36] so I think it never finishes the image install [15:37] *boots to the os of the ephemeral image (I think) [15:37] then nothing else happens, it just sits until it hit 40mins and then goes into failed deploy status in maas [15:38] roaksoax: I feel like I just need to get better insight into whats going on here somehow [15:41] geh, so much ambiguity [15:41] bdx: well, in the UI, check the 'Events' tab and pastbin what you have there [15:42] roaksoax: https://imgur.com/a/w1dpi [15:43] bdx: can you view more ? [15:44] yeah, https://imgur.com/a/PiFGy [15:45] 1046 machine events in the past 1 days [15:45] quite a bit of activity [15:46] bdx: that is telling me that the machine attempted to PXE boot [15:46] and probably got stuck there [15:46] bdx: can you get a full log since you started the installing [15:46] ok [15:47] yeah, where can I get that ? [15:48] like, you want the cloud-init.log from the node? [15:48] the nodes are prompting me for a password when I try to ssh in [15:49] oh, you just want the full event log [15:50] roaksoax: would an "Export Event Activity" button be useful in the events page you think? [15:51] bdx: maas events query hostname= level=INFO limit=200 [15:51] oh sweet [15:52] bdx: some of our systems hang during pxe right after loading boot-initrd. it only happens on systems that are in a Cisco ACI network. I was wondering if you have an ACI network also? [15:52] shoot, I'm unfamiliar with ACI [15:53] I see its some sort of Application Centric Infrastructure? [15:53] lol [15:54] let me see If I can figure this out [15:54] bdx: ok no biggy was just wondering. ACI is Cisco's SDN. There are no firewalls in the data path and its intermittent. If we reboot it a few times it will usually proceed and deploy. [15:56] ahh, nice, ok, yeah we just purchased some used cisco 10G switches from like 5 years back, I highly doubt they are using ACI, but I just dont know [15:57] roaksoax: http://paste.ubuntu.com/26065391/ [15:57] am i missing something [15:58] bdx: use just the hostname [15:58] bdx: no need for the domain [15:58] bdx: and, are you logged into the cli ? [15:58] oh, I need to login before the admin command will become available? [15:59] maas login http://:5240/MAAS [15:59] got it [15:59] thx [16:00] roaksoax: http://paste.ubuntu.com/26065418/ [16:00] that is just great [16:00] I guess I need to dive into the api/cli a bit more === disposable3 is now known as disposable2 [16:18] bdx: http://pastebin.ubuntu.com/26065507/ this is telling me that the machine powered on, started to PXE boot, but then it never got the kernel nor the initrd [16:18] and got stcuk there [16:19] So, the machine is booted to the ephemeral image it seems now [16:19] $ ssh ubuntu@10.10.20.3 [16:19] ubuntu@10.10.20.3's password: [16:20] if I look at the ipmi viewer, its sitting at the ubuntu login screen [16:22] bdx: maybe something failed and falled back into the local disk ? [16:22] bdx: can you fdisk -l ? [16:23] "but then it never got the kernel nor the initrd" - your thinking the phases 1 of deployment failed and it just booted to local [16:23] yeah [16:23] I cant login [16:23] it doesnt seem to have my key, and is prompting me for pw [16:23] yeah that could be it, I would attempt to try to re-deploy [16:23] bdx: and look at the machine console via the ipmi viewer [16:24] to see what it is doing [16:24] roaksoax: right Ive done this a binch now [16:24] blah, *bunch [16:25] and cant figure it out [16:25] its happening to all of my nodes [16:26] bdx: fwiw, i deployed machines using maas from the snap just fine yuesterday [16:27] ok [16:27] bdx: tail the logs in /var/snap/maas/common/log/ regiond.log rackd.log if you see anything aspecific [16:29] roaksoax: http://paste.ubuntu.com/26065565/ [16:29] seeing that in rackd.log [16:34] ahah [16:34] got it [16:36] http://paste.ubuntu.com/26065581/ [16:36] roaksoax: I updated the bug with ^ [16:40] roaksoax: do you want two separate bugs for those^? [16:49] I wonder why you didnt hit that [16:50] roaksoax:^ [16:52] bdx: yeah, the port/socket thing is not an issue though [16:52] the other one is what concerns me [16:52] bdx: i'll re-deploy another environment [16:52] 1 sec [17:04] bdx: while i look into this, what does your proxy.log say ? [17:05] http://paste.ubuntu.com/26065728/ [17:06] $ sudo cat /var/snap/maas/common/log/proxy/* | http://paste.ubuntu.com/26065733/ [17:09] bdx: strange, squid doesn't run for me in the snap [17:11] huh [17:11] that is odd [17:11] but that's an issue in my machine [17:11] roaksoax: i'm running region+rack [17:11] mode [17:11] not sure if that has anything to do with anything [17:13] oh, like you already had something running on the port 8000 so it didnt start [17:13] or something [17:19] bdx: yeah, my issue is only on ym machines for some reason, it works on other machines i have [17:19] bdx: by anyway, re-deploying my maas form the snap [17:19] bdx: to test [17:21] ok, thank you [17:37] bdx: yeah, that's the issue... strange [17:38] something changed in curtin? [17:38] refreshing to not be operator error for once :) [17:39] bdx: maybe, but the snap shouldn't have been affected. I'm double checking now === frankban is now known as frankban|afk [18:15] bdx: figured out what the issue is, building a new snap and will test it [18:22] awesome [18:24] roaksoax: what did you find to be at the root of this? [18:36] bdx: bug in the snap creation when we transitioned to not using upsteram curtin [18:40] ahh [18:41] thanks for help with this [19:42] Bug #1734980 opened: [2.3] Packaging doesn't remove /dev/shm/squid-cf__metadata.shm [21:20] vogelc: have you tried a different kernel ? [21:20] e.g. a hwe-16.04 kernel