[02:27] guess I’ll answer my own question. Apparently the HTTP proxy plays a part in the IPMI discovery. Opened TCP 8000 and enlistment was able to determine the power type to be IPMI 2.0 [02:29] that's very odd [02:29] trial and error got me there. [02:29] roaksoax, possible? ^^ [02:29] but I suppose my trials could have had an error :D [02:30] watched dropped packets on MAAS, opened things up until enlistement revealed the power type as it had with no firewall [05:41] benlake: maas installs the ipmitool package on the machine. if it can't use the http proxy, it can't install that package, and thus not determine the power type. === frankban|afk is now known as frankban [08:12] morning [08:47] good morning cnf [08:47] \o [09:51] Hi All [09:51] i am using maas 1.9 version [09:52] is there support to deploy 16.04 for any provision [13:23] /win 6 [13:36] BjornT: ah, that makes sense. [13:52] sanjay: you can deploy 16.04 with 1.9 [13:52] sanjay: you cannot commission with 16.04 in 1.9 [14:11] i have mass 2.1.5 [14:11] I can commission nodes [14:12] but while deploying at last moment its shows error of grub rescue> [14:12] node is at final stage to get delpoyed but if grub rescue doesnt come [14:14] sanjay: that seems like a deployment failure [14:14] sanjay: grab the installation log from the UI [14:14] on the node that failed [14:14] go to the bootom part [14:14] where the machine output is [14:14] and on the dropdown select installation log [14:14] and pastebin that [14:16] error: attempt to read or write outside of disk 'hd0' [14:20] sanjay: can you pastebin it all please ? [14:20] i can see the outputsummary logs from webui [14:21] how to use pastebin [14:23] https://pastebin.com/xWCE4PYT [14:24] sanjay: on the dropdown menu, what options do you have on that sectio n? [15:10] Bug #1693255 opened: [2.2, UI] Cannot add an interface when the machine is "NEW' [15:10] Bug #1693256 opened: [2.2, UI] Cannot add an interface when the machine is "NEW' [15:27] after setting boot disk for a machine in maas [15:27] how do I apply that? the machine still shows sda as the boot disk, even though I got a 200 back from the API :/ [15:29] well let me put it this way, it still shows up as the install disk in the storage tab. do I need to combine boot disk with something else to change the install disk? [15:35] I’m very new to MAAS, but I get the impression any changes such as the one you speak are Deploy time implemented only. [15:36] I don’t think it is designed to make changes post deploy [15:37] one more thing [15:38] does we need to have atleast 2 disks on server [15:38] or with 1 disk also OS can be deployed [15:42] one disk is fine. what is your thinking around needing 2 disks? [15:58] no just asking [15:58] do we need to do any partition [15:58] it creates on GPT partition and other part-2 partion with full disk [16:06] if you are using EFI, yes [16:22] sanjay: before you deploy, you can mess with the partitioning scheme on the node [16:23] unfortunately there doesn’t appear to be a way to save that scheme and apply it to other machines. I assume this isn’t a new thought and someone is thinking about custom layouts, eventually. [16:39] benlake, we're working on that right now in fact [16:39] the first iteration will be a dump-to-yaml and upload-yaml [16:40] look at that, heroes of the people [16:40] that’s a good start [16:40] removes all the UI work [16:42] we are also doing something interesting for our use case, which is grouping machines into buckets of "identical" config [16:42] so you'll be able to easily configure a set of machines with the same layout [16:42] and by identical config, layout I mean # and type of NICs, drives, and cpu/core and RAM count [16:45] kiko: accepted :P [16:45] thanks! [16:46] :P [17:11] Bug #1693299 opened: MAAS allows any image name to be uploaded by only known images can be deployed [17:17] Bug #1693299 changed: MAAS allows any image name to be uploaded by only known images can be deployed [17:20] Bug #1693299 opened: MAAS allows any image name to be uploaded by only known images can be deployed [17:25] Guys i have deployed 4 nodes..i created one /boot partition and other / root partition and started deployed and it worked for me [17:25] thanks all for your support [17:26] hope this can be helpful to other people who are facing same error [17:27] kiko: +1 to identical layouts [17:27] benlake, thanks -- I think it's going to be really useful [17:28] hey guys, is it possible, via MaaS -> Interfaces, to create OpenvSwitch Bridges? And not only Linux Bridges...? [17:48] kiko, i think this is also on the map but, to be sure, please allow (storage) percentages when changing a layout [17:52] currently sizes need to be in, gulp, bytes [17:57] pmatulis, yes indeed [17:57] ThiagoCMC, not currently [17:57] Hmm... Ok! [18:04] does anyone can give exact layout in % wise when doing partition manually [18:12] Hi everyone! I see that maas region controller and maas rack controller DNS registers are under ".maas" domain. Is is possible to change these domains? Can be different for rack controller and region controller? Finally, could some one guide me and tell me how to change this? [18:20] nobody? [18:24] you can add more domains [18:24] and move the nodes to it, before deploying [18:30] Mass dhcp not running [18:30] can anyone please let me know how to start it [18:46] reboot [18:46] =P [18:53] ThiagoCMC: I created more domains , I also deployed new servers and I know how to assign them to the new domains. However, I can't do this with maas rack controller and maas region controller [19:18] roaksoax: seems root_disk is ignored on custom image? [19:22] benlake: https://bugs.launchpad.net/maas/+bug/1418044/comments/3 -> in your comment there, what version of MAAS are you using? What was set in /etc/maas/regiond.conf and what was set in /etc/maas/rackd.conf ? [19:23] xygnal what is the custom image you are trying to deploy ? === mup_ is now known as mup [19:47] a centos7 cloud image we have made some tweaks on [20:30] Guys, at the end of a deployment, I'm seeing the following message: "cloud-init[XXXX]: Can not apply stage final, no datasource found!" [20:30] What to do? [20:31] I'm running MaaS Next [20:31] It was working this morning but, I updated the box because a new package appeared on MaaS Next PPA... [20:31] xygnal: at the moment we dont support any storage config for centos, I think that's the culprit. [20:32] xygnal: you could customize the image curtin hooks to install where you need them to [20:32] ThiagoCMC: has it caused a failed deployment ? [20:32] I was stucked... [20:32] I rebooted the node, boom! It worked! [20:33] ThiagoCMC: strange, it would be interesting to see the full cloud-init log if you can grab that [20:33] Sure, I'll collect it and post on pastebin [20:33] ThiagoCMC: but final has had tiny bug fixes not at all related to metadata server or anything of the sorts [20:33] Hmm... I see [20:33] which file should I collect? [20:34] ThiagoCMC: /var/log/cloud-init*.log from the ephemeral env if you can, and also /var/log/maas/regiond.log and maas.log [20:34] ok [20:47] roaksoax, figured out! For some reason, MaaS picked up the wrong subnet for a fabric... So, it gave two IPs from the same subnet, to two different segments, which was causing problems with the communication between cloud-init and the metadata server. [20:49] roaxsoak: is that neccesary? can customs or centos not use root_device? [20:50] roaxsoak ok i see. any example code? [21:23] roaxsoax: going to need a little more info on using curtain [21:24] roaxsoax: for that purpose [21:26] using block-meta? [21:29] xygnal: so IIRC< curtin will pick up the first device it finds in the list [21:29] Bug #1693358 opened: VLAN name mismatch between web UI and API [21:30] xygnal: and on the OS' where we dont configure the underlying store, like centos, we dont pass along any storage config because that could actually break the deployment of things [21:31] do i neee to pre-label it so it knows to pick it up? [21:31] label=root or such? [21:32] xygnal: i need to step out right now. I'll try to come back later, but otherwise we can pick that up tomorrow [21:34] ok. [22:00] where does my curtin_userdata_custom need to be, region or rack server? [23:29] Bug # changed: 1687420, 1688060, 1689288, 1689838, 1690144