[00:51] catbus1: hate it when that happens. ;-) I (had a bad Ethernet cable cause me a half-day triage of what I thought was a MAAS bug the other day!) [01:21] Bug #1677933 opened: [2.2] Adding a virsh pod should add machines with their given names [01:21] Bug #1677936 opened: [2.2] Failure while handling notification to 'event' [01:39] Bug #1677940 opened: [2.2b4] XML/YAML output is huge and increases machine page load time [01:48] Bug #1677940 changed: [2.2b4] XML/YAML output is huge and increases machine page load time [01:51] Bug #1677940 opened: [2.2b4] XML/YAML output is huge and increases machine page load time [04:27] Bug #1660169 changed: Upgrade resulted in a rack-controller with server BMC info [09:59] hello there [09:59] how can i change the default timezone of the deployed systems [10:07] BlackDex, is ntp not getting the right timezone? [10:16] Bug #1678038 opened: [2.2 UI] Edit button on Machine Summary tab of Node details page is not aligned [10:26] brendand: doesn't look like it [10:26] but i found that juju sets all to UTC [10:27] so i think that is the case here [10:27] but i'm not sure [12:36] Hi All [12:36] is there any way that i can add a custom RHEL image to MAAS? [12:38] BlackDex: is your maas server in X timezome, and when you deploy the machines dont end up with the same timezone as your MAAS ? [12:38] BlackDex: also, have you tried deploying without juju to see the outcome [12:39] roaksoax, allenap said that deployed systems will be in UTC, dunno if that's regardless of the maas servers timezone [12:39] Chaitu: you can add images to maas similar like: maas admin boot-resources create name=custom/rhelx title="RHEL X" architecture=amd64/generic content@=yourrhelimage.tgz filetype=ddtgz [12:40] roaksoax: haven't tried that [12:40] Chaitu: but it is likely to fail deploy as it would need to be customized for MAAS [12:40] but yes that is basicly the issue [12:40] and we only provide custom images for that to customers atm [12:51] the important thing with custom images is, that if you want maas to manage everything you need to install cloud-init [12:51] that should make a custom image work [12:52] i created a custom ubuntu-desktop image, without cloud-init, and that works fine, it deploy's it and all. [12:52] But that is probably because it is a ubuntu image [12:52] don't know how it works for centos/rhel/fedora images [13:01] Bug #1673655 changed: deploy of a custom image failed [13:07] Bug #1673655 opened: deploy of a custom image failed [13:08] roaksoax: Hi, Thanks for the reply ,But i tried build the custom image ,but could not install maas-image-builder [13:09] roaksoax: Is there any utility where i can create custom images for MAAS? [13:13] Chaitu: at the moment we are only providing maas-image-builder to our customers only [13:13] so there's no longer a public utility [13:13] BlackDex: that's only for ubuntu images though :) [13:13] BlackDex: and for desktop images there's lp:curtinator [13:16] roaksoax: Is the support pricing for each MAAS server or per bare metal server? [13:22] Bug # changed: 1633726, 1640301, 1642040, 1650481, 1664384, 1669248, 1669833, 1671048, 1671891, 1672354, 1672837, 1672841, 1673378, 1673655, 1673724, 1674788, 1677005 [13:22] hi, would it be possible to make #1677507 public? [13:25] kklimonda: it is fixed, fix will be in 2.1.6 (next week) or in 2.2.0b5/or 2.2.0rc1 [13:28] kklimonda: fixed [13:28] (the private stfuf) [13:28] roaksoax: thanks, I just found the commit :) [13:29] (after learning that it was already in bzr) [13:30] is there any mechanism within MaaS to limit DNS zone transfers to a few IP's or a specific subnet? [13:30] I know that this can be done directly through bind, but was not sure whether the configuration would be overwritten by MaaS [13:31] tychicus: there's no such mechanism [13:31] lamont ^^ advice ? [13:32] roaksoax: thanks, just wanting to make sure I wasn't missing something in the Documentation [13:32] lamont may give some advice though [13:34] tychicus: we don't currently have a way for you to provide restrictions on axfr/ixfr at the zone level. Let me see how bad it is at the global level [13:34] Bug #1677573 opened: [2.2] Static addresses can be allocated in dynamic range [13:34] tychicus: having said that, MAAS does have "things" in the global options... now I just need to figure out how those get set, and I'll have a real answer for you [13:35] lamont: thanks, a global setting would be great [13:37] tychicus: it turns out that if you configure any of "allow-query, allow-recursion, and allow-query-cache" in /etc/bind/named.conf.options, maas will honor them. [13:38] tychicus: which means, if you already have maas installed and running, you need to migrate them back to .options from /etc/bind/maas/named.conf.options.inside.maas, configure them, and then restart [13:38] which may be worthy of a low-pri bug report [13:43] lamot: I'll give it a try and report back [13:44] happy to submit a bug report as well [13:44] ta [13:44] thanks [13:44] I know it works well when BIND is preinstalled and restricted. not so sure about the after-the-fact handling [13:46] good to know, I may not get to it until next week, but I'll do my best to document and report the experience [13:46] Bug #1677573 changed: [2.2] Static addresses can be allocated in dynamic range [13:52] Bug #1677573 opened: [2.2] Static addresses can be allocated in dynamic range === gimmic_ is now known as gimmic [14:57] hey guys, MaaS add the following line to VLAN interfaces: "mtu 1500" but, looks wrong... I think it applies only to the physical interfaces, not to the VLAN interfaces. [14:57] The following error appear: "/etc/network/if-pre-up.d/mtuipv6: line 9: /sys/class/net/eno1.100/mtu: No such file or directory" [14:58] when I try to "ifup eno1.100" [15:10] Bug #1678188 opened: [2.2r5898] 'View full history' for machine events does nothing [15:13] ThiagoCMC: did you set the MTU on the interface as high as on the vlan? [15:17] I didn't set any mtu, it is a fresh installed maas, I'm just using the defaults. [15:17] blake_r: (or others, but istr seeing a reply somewhere from you on this with an earlier version) What's the current best recommendation for juju2+maas wrt the bootstrap/controller node? kvm? lxd? [16:13] Bug #1678215 opened: user role , admin role [16:19] ? [16:43] Bug #1678231 opened: [2.1.5] node marked broken then fixed retains auto assigned IP in ready state [16:43] Bug #1678236 opened: "Requested reserved range conflicts with an existing range", but no conflict shown === frankban is now known as frankban|afk [18:15] mpontillo: Hey one more quick question - My setup again is the region controller in AWS (using local DB now) and a rack controller physically on site. I have separated the production and BMC networks, and MAAS is not serving DHCP for the BMC network. MAAS has detected all the correct BMC information, but still reports no rack controller can access the host [18:16] when at layer 3, my rack controller can absolutely access the BMC address [18:17] does MAAS *have* to also provide dhcp for the BMC network for this to work? [18:17] cmd_pancakes: not provide DHCP, but MAAS does prefer to be able to be on the same network as the BMCs, so it can determine which rack would be the best one to communicate with the BMCs [18:18] cmd_pancakes: so probably something is going wrong with the logic to try to figure that out. [18:18] blake_r: ^ FYI [18:18] cool...yeah i've tried to add the new subnet to the same fabric, but that still didn't do the trick [18:19] cmd_pancakes: MAAS actively tried to be able to query your BMC's [18:19] most likely the rack controller cannot authenticate to your BMC [18:20] blake_r: The new host successfully PXE'ed the first time, and the initial MAAS record shows the correct IP address...I assume the initial boot image set the maas user and password for the BMC interface? [18:20] i can try it manually from the host with the creds on file [18:22] cmd_pancakes: you might try going to a node details page and clicking "check power"; when MAAS does a power query, if it cannot find an on-link rack that has direct BMC access, it will try to determine which racks can route to the BMC [18:23] cmd_pancakes: while you're there, ensure the credentials are correct, as blake_r said =) [18:23] mpontillo: ohhh sorry sorry, i think i know the issue...by default Dells come with IPMI over lan disabled, i bet thats it [18:23] yeah when i tried to do it manually it timed out with ipmi 1.5 and 2.0 [18:24] cmd_pancakes: yeah there is a bug about MAAS turning that on during enlistment [18:24] cmd_pancakes: https://bugs.launchpad.net/maas/+bug/1664822 [18:25] ok gotcha...in the past i rolled a lot of this automation myself and used dell's racadm for enabling it [18:25] yep, cool, we're on the same page :) [18:27] blake_r: so just to recap since i've played around with a lot of settings to try and get this to work, i *don't* need the BMC subnet configured on a rack controller fabric or vlan for this to work? as long as the correct ipmi settings are detected during enlisting and accessible over layer 3, MAAS is trying to just autodetect with those settings [18:27] cmd_pancakes: correct [18:27] cmd_pancakes: just need layer3 access [18:28] cmd_pancakes: and the abilitty to authentication [18:28] excellent, thanks for the help! and mpontillo [18:28] cool deal [18:33] yeah ipmi over lan was the issue...works great now [19:17] Hey guys, how to add a CNAME in MaaS' DNS? I have: "eth1.100.select-fish.maas", and I want a CNAME: "openstack-1.admin.maas -> eth1.100.select-fish.maas". [19:19] Is it possible to add CNAME using MaaS CLI? I can't find a way to do this via its Web UI. [19:21] lamont: ^^ [19:22] Looks like that this is an old question as well: http://askubuntu.com/questions/775123/is-it-possible-to-add-custom-a-or-cname-records-into-zone-controlled-by-maas-1-9 =P [19:22] No replies [19:24] Hmm... https://www.lucaswilliams.net/index.php/2016/10/02/maas-2-0-and-dns/ [19:38] Bug #1678298 opened: [2.2] Action button for "Map subnet" is misleading [20:18] ThiagoCMC: yes: maas SESSION dnsresource-records create fqdn=LHS.DOM.AIN rrtype=CNAME rrdata=TAR.GET.FQDN. <-- note the trailing dot... [20:19] Sounds awesome! Thank you! [20:19] Looks like that MaaS CLI is way much more powerful than its Web UI... =P [20:20] the goal is to bring the webui into "largely parity, ish" -- reality is that most large installs are scripting everyting and hence using the CLI, so things tend to get added there first, and then added to the webui where it makes sense. [20:21] dunno if there's a bug report in launchpad asking for DNS ability via the webui, but you're certainly welcome to file such a thing [20:21] Fair enough... =) [20:30] lamont: there is [20:30] the bug exists [20:30] ThiagoCMC: yes the CLI does absolutely everything the UI does + obviously more [20:31] there's a lot of things we dont expose over the API/CLI [20:31] Sounds cool! I never used maas cli before, sounds like that I'll have an awesome weekend! :-D [21:14] Hi, I tried to create a partition on sda, it shows sda-part1, but after I saved the configuration, it shows up as sda-part2. [21:35] catbus1: sda-part1 is probably the efi partition ? [21:37] roaksoax: deploying failed and showed error:attempt to read or write outside 'hd0' and entered the grub rescue mode [21:50] roaksoax: maybe related to this bug https://bugs.launchpad.net/ubuntu/+source/ubiquity/+bug/1284196 [21:51] it was the only machine that has 4tb partition and it was booting in efi mode. [22:20] Bug #1678323 opened: [2.1.5] Power parameters stopped working for UCSC-C220-M4L Cisco system [22:32] Bug #1678323 changed: [2.1.5] Power parameters stopped working for UCSC-C220-M4L Cisco system [22:38] Bug #1678323 opened: [2.1.5] Power parameters stopped working for UCSC-C220-M4L Cisco system [22:45] roaksoax: created a 2TB partition and the deploy went fine. [22:45] fyi. [23:29] Bug #1678333 opened: [2.2] UI regressions on machine listing page