| shewless | Hi there. got a new install of maas 2.0 on ubuntu 16.04. Setup a DHCP network/subnet on my vlan but I'm noticing that maas-dhcpd is not starting. I see two "problems" in syslog | 01:35 |
|---|---|---|
| shewless | 1: "no subnet declartion for ens3.200" - but I declared a subnet on this vlan in maas. | 01:35 |
| shewless | 2: "Not configured to listen on any interfaces" - Hmm. not sure why | 01:36 |
| shewless | hmm.. can I not use a tagged vlan for pxe? | 01:43 |
| shewless | I think I may have found the issue. I named the vlan interface "vlan200" instead of the standard "ens3.200". Looks like maas made some assumption that my vlan interface is ens3.200. I changed my vlan interface name and it seems to be working well now., | 02:05 |
| shewless | Hmm.. well that change allowed my system to pxe boot, but now I'm staring at "maas-enlist" IP-Config: no response after X secs - giving up | 02:10 |
| shewless | next problem.. what is multipath? | 02:30 |
| shewless | sigh.. actually I think it just can't mount the iscsi drive because the IP-Config timed out | 02:39 |
| shewless | looks like it's trying to configure the physical interface and it's not trying to configure the vlan.. so maybe pxe on a tagged vlan just isn't supported? Can anyone verify? | 02:39 |
| === frankban|afk is now known as frankban | ||
| martin123 | hi all.. quick question: I've looked around in the docs and googled for it but I can't seem to find much... I need to move a maas (1.6) controller from old hardware to newer hardware and was hoping to do an upgrade to maas 2 as well... anyone know of good docs or have any tips? There appears to be no export/import functionality as far as I know. | 09:41 |
| roaksoax | martin123: i'd upgrade to 2.0, then fully backup the DB and backup /var/lib/maas/secret | 09:50 |
| roaksoax | martin123: and reapply the backedup db to the new system | 09:50 |
| roaksoax | and restore /var/lib/maas/secret | 09:50 |
| roaksoax | and that should be it | 09:50 |
| martin123 | roaksoax: ok, sounds logical. Thanks, I'll try that :) | 10:21 |
| === Guest20905 is now known as CyberJacob | ||
| mup | Bug #1616417 opened: [2.0] The IP of a Machine's BMC is not listed on the subnet details page <MAAS:Triaged> <MAAS 2.0:Triaged> <MAAS trunk:Triaged> <https://launchpad.net/bugs/1616417> | 11:04 |
| shewless | hello. anyone active here? Just wondering if tagged vlans are supported for the pxe network | 14:57 |
| roaksoax | shewless: have you configured your hardware to pxe over a tagged vlan ? | 15:00 |
| roaksoax | shewless: we don't test that scenario, but if the hardware firmware were to support that, I don't see why that wouldn't work | 15:01 |
| roaksoax | you'd just need MAAS available on such vlan | 15:01 |
| shewless | roaksoax: yes I have. I can see the dhcp request and offer works but during the commissioning I can see it's trying to assign the IP to the pysical nic instead of the vlan.. it all goes downhill from there :) | 15:01 |
| shewless | roaksoax: I'm pretty sure it doesn't work.. just wanted to confirm. For now I've removed the tag | 15:02 |
| shewless | roaksoax I have another problem maybe you can help me with :) | 15:02 |
| shewless | I'm deploying a server (16.04) and I see this message after it's done most of the imaging and is just doing the final reboot (WARN: No MBR magic, treating disk as raw). | 15:03 |
| shewless | I have 2 servers configured the same. works on 1 and not on the other | 15:03 |
| shewless | I am trying to install the OS on a USB stick.. maybe that's relavent | 15:04 |
| roaksoax | shewless: ah yes, the commissioning environment wont configured a tagged NIC on the machine | 15:05 |
| shewless | roaksoax: that explains that then. I'll just avoid the tag on the pxe network then | 15:05 |
| roaksoax | shewless: on your second issue, did you setup your usb disk as the boot disk ? do you want your usb disk be the boot disk ? | 15:06 |
| shewless | roaksoax: Okay.. I now have 3 servers configured the same. On two of them I can deploy just fine. On one I get that error about MBR magic. I tried to replace the USB stick and I also tried to dd if=/dev/zero of=/dev/sdb bs=512 count=1 to clear out the MBR.. just in case that would help | 15:07 |
| shewless | roaksoax: Yes and Yes | 15:07 |
| shewless | it works on 2/3 servers.. weird | 15:07 |
| roaksoax | shewless: strange. Try to reocmmission your non-working machine and re-do the storage config and see if that makes any effects ? | 15:08 |
| roaksoax | s/makes/takes | 15:08 |
| shewless | roaksoax: tried that already twice.. | 15:09 |
| shewless | in the rsyslog I see this.. not sure if it's relavent: ureadahead: Unable to obtain rotationalness for device 0:22: No such file or directory | 15:09 |
| roaksoax | never seen such message | 15:09 |
| roaksoax | shewless: what if you change the boot disk ? | 15:10 |
| roaksoax | shewless: and change it back to what you want it to be | 15:10 |
| shewless | roaksoax: it works if I use the hard drives | 15:10 |
| shewless | ..and then doesn't work when I switch it back to what I want | 15:10 |
| roaksoax | strange | 15:14 |
| roaksoax | i'd suggest you file a bug | 15:14 |
| shewless | roaksoak.. thinsk I fibured it out.. will let you know after I test | 15:18 |
| shewless | roaksoax: I had the drive boot order in the bios set backwards for that one device | 15:31 |
| shewless | that somehow caused this problem | 15:31 |
| roaksoax | strange indeed | 15:32 |
| shewless | roaksoax: Is there a way to change the default use when deploying a node? I see it's ubuntu even though I'm deploying as a different user in maas | 15:35 |
| roaksoax | shewless: no, not yet | 15:35 |
| === frankban is now known as frankban|afk | ||
| === mup_ is now known as mup | ||
| === mup_ is now known as mup | ||
| kiko | shewless, there's the argument that it should always be ubuntu, btw | 16:29 |
| kiko | because this is meant to be an admin user | 16:29 |
| kiko | not an actual machine user | 16:29 |
| shewless | kiko: fair enough. Not that hard to work with. | 16:32 |
| kiko | shewless, our thinking is that for any simple case, it's trivial to useradd once you have a user on the machine | 16:33 |
| kiko | shewless, and for any complex case, the ubuntu user is really just being used to bootstrap automation | 16:33 |
| kiko | it's a bit of a question mark that the user should be called "ubuntu" and "centos" respectively | 16:33 |
| kiko | roaksoax, is the rhel user "rhel"? | 16:33 |
| kiko | (i.e. we could use maas or maasadmin or maasuser) | 16:34 |
| shewless | kiko: just call it "maas" then :) | 16:34 |
| shewless | kiko.. yeah that might make more sense | 16:34 |
| shewless | shewless: I think it also might make sense to make that user name the same as the deployers user name | 16:34 |
| shewless | or at least to provide that as an option | 16:35 |
| kiko | shewless, the latter makes it harder for the handoff-to-automation use case | 16:35 |
| shewless | kiko: not in my situation :) | 16:35 |
| kiko | shewless, can you explain? | 16:36 |
| roaksoax | kiko: root iirc | 16:37 |
| shewless | kiko: actually having thought about it I suppose it doesn't matter. Just that I'm running some ansible deployment scripts and I generally run those as the same user as my maas user | 16:37 |
| shewless | which is not ubuntu | 16:38 |
| shewless | (and not root) | 16:38 |
| kiko | shewless, do you use maansible? | 16:38 |
| * roaksoax otp | 16:38 | |
| shewless | kiko: I wasn't aware of such a beast. Definitely was in the plans to integrate the two at some point | 16:39 |
| kiko | roaksoax, and not with me it seems? :) | 16:43 |
| kiko | shewless, I've never used it but know of it: https://github.com/waldemarbautista/maansible | 16:44 |
| shewless | kiko: yeah I might play with that at some point. For now I'm okay with a two step process of deploying and ansibling. | 16:44 |
| shewless | with teh ubuntu user :) | 16:44 |
| roaksoax | kiko: sorry, got stuck in another meeting | 16:45 |
| kiko | no problem but my time is running out | 16:47 |
| shewless | kiko, roaksoax: I'm having a heck of a time deploying on a KVM machine. I got the virsh power management working, commissioning works, but it fails on deployment | 17:00 |
| shewless | http://paste.ubuntu.com/23085690/ | 17:00 |
| shewless | that's the rsyslog.. any ideas? | 17:00 |
| kiko | Aug 24 12:46:56 wtllab-guest-1 cloud-init[2310]: An error occured handling 'vda': OSError - [Errno 6] No such device or address: '/dev/vda2' | 17:03 |
| kiko | Aug 24 12:46:56 wtllab-guest-1 cloud-init[2310]: [Errno 6] No such device or address: '/dev/vda2' | 17:03 |
| kiko | Aug 24 12:46:56 wtllab-guest-1 cloud-init[2310]: Installation failed with exception: Unexpected error while running command. | 17:03 |
| kiko | Aug 24 12:46:56 wtllab-guest-1 cloud-init[2310]: Command: ['curtin', 'block-meta', 'custom'] | 17:03 |
| kiko | Aug 24 12:46:56 wtllab-guest-1 cloud-init[2310]: Exit code: 3 | 17:03 |
| kiko | Aug 24 12:46:56 wtllab-guest-1 cloud-init[2310]: Reason: - | 17:03 |
| kiko | Aug 24 12:46:56 wtllab-guest-1 cloud-init[2310]: Stdout: b"An error occured handling 'vda': OSError - [Errno 6] No such device or address: '/dev/vda2'\n[Errno 6] No such device or address: '/dev/vda2'\n" | 17:03 |
| kiko | shewless, you don't have a vda2 on that VM? | 17:03 |
| shewless | kiko: I don't think so (not on purpose). the config is only for vda and vda1 I believe | 17:04 |
| shewless | kiko: maas just has "vda" and vda-part1. I didn't mess with the disk config at all | 17:05 |
| kiko | shewless, well, that's the only thing I can spot from the logs | 17:05 |
| kiko | why is MAAS trying to talk to vda2? | 17:05 |
| kiko | is it a UEFI thing, i.e. vda1 will be the UEFI partition? | 17:05 |
| kiko | shewless, is the deployment actually failing, btw? is all of this prior to running curtin? | 17:06 |
| shewless | kiko: maas is telling me the deployment failed. the console looks like it got pretty far. | 17:07 |
| shewless | I'll try and create a different partition maybe | 17:07 |
| mup | Bug #1616552 opened: [IPv6] ipv6-only machine has region's ipv4 address in resolv.conf <MAAS:New for lamont> <https://launchpad.net/bugs/1616552> | 17:20 |
| === mup_ is now known as mup | ||
| valeech | Hello. I am running maas 2.0rc4. Whenever I try to deploy a machine with xenial it gets stuck deploying. It seems to install and reboot ok, but then nothing happens after the reboot. I can deploy trusty all day long without issue. Any pointers where to start troubleshooting? I tried commissioning with both trusty and xenial and nothing changes. I am using the xenial hwe-x kernal in all situations. | 19:20 |
| jpr1 | trying out maas 2.0 on xenial. having a bit of trouble getting my dhcp requests to get a reply | 19:34 |
| jpr1 | The documented "nodes accept-all" doesn't seem to be an option https://maas.ubuntu.com/docs/nodes.html#automatic-discovery | 19:35 |
| jpr1 | I see the dhcp requests on the maas server via tcpdump but there are no corresponding replies | 19:36 |
| jpr1 | (note that I am in liberty openstack virtual environment but I do have the subnets dhcp service turned off) | 19:37 |
| jpr1 | any thoughts? | 19:37 |
| === mup_ is now known as mup | ||
| === mup_ is now known as mup | ||
| sunny_ | can anyone give me pointer for openstack-charmer channel as /JOIN openstack-charmers is not working for me | 20:32 |
| marcoceppi | sunny_: /join #openstack-charms | 20:34 |
| sunny_ | Thanks marcoceppi | 20:38 |
| Dougi | hi all, im new to openstack, but im very interesting to learn, so i have a ubuntu maas setup with a controller and 4 nodes deployed, but i dont know what next... and i kind of dont find any good dokumentation. can someone tell me how to find any good dokumentation on how to get juju installed correctly? | 20:39 |
| marcoceppi | Dougi: #juju might be a good place to start, but you can also check out https://jujucharms.com/docs/stable/getting-started-general | 20:43 |
| marcoceppi | Dougi: you install Juju on your machine, not on maas or anything directly | 20:43 |
| marcoceppi | Dougi: Then you tell juju about your current MAAS controller, which enables you to create a Juju controller, which you can then deploy OpenStack with Juju | 20:43 |
| Dougi | aha.... | 20:44 |
| Dougi | oki... thats make more sense.... | 20:44 |
| marcoceppi | Dougi: here's the MAAS documentation to configure Juju and MAAS https://jujucharms.com/docs/stable/clouds-maas | 20:45 |
| Dougi | thanks marcoceppi :) | 20:45 |
| Dougi | marcoceppi: there is nowhere in any documentation that says that i need to use my machine :) i tought allways that i had to installl juju at the controller or maybe a node so that i could start the webgui :) | 20:50 |
| tester | hi | 20:50 |
| tester | is anyone out there? | 20:51 |
| kiko | maybe | 20:51 |
| kiko | if you have some money to send us for instance then we are all likely here | 20:51 |
| tester | haha i see | 20:52 |
| tester | do you use bitcoin? | 20:52 |
| kiko | tester, no, mostly MAAS. what about you? | 21:02 |
| Dougi | marcoceppi: is there any maps or something that shows how evrything is connected to eachother? | 21:07 |
| marcoceppi | Dougi: not that I'm aware of | 21:08 |
| tester | kiko you mean you have weapons of maas deployment | 21:09 |
| tester | anyone know how to deploy rhel in maas? | 21:09 |
| kiko | tester, centos works out of the box | 21:10 |
| kiko | tester, rhel is delivered to paying customers as we can't redistribute it publicly | 21:10 |
| kiko | centos is supported out of the box in both 1.9 and 2.0 | 21:10 |
| tester | okay | 21:12 |
| tester | i have redhat images and licences can i create my own custom image? | 21:12 |
| kiko | tester, if you are a paying customer yes | 21:14 |
| kiko | we give you all the tooling | 21:14 |
| tester | does the tooling support ppcle arch? | 21:16 |
| tester | the free centos image suports only amd | 21:16 |
| tester | *amd64* | 21:17 |
| Dougi | marcoceppi: so i can not run maas-controller and juju at the same server... but can i run juju at a server outside of my "mass network" that controll my maas cloud? is that correct? | 21:23 |
| kiko | tester, good question -- you are on ppcle | 21:25 |
| tester | yes i am | 21:26 |
| tester | @kiko i am in need of boot images for rhel6 & 7, suse 11 & 12 (ppcle, s390x and amd64) | 21:30 |
| kiko | tester, ppcle is power8, right? | 21:31 |
| tester | @kiko maas deploys ubuntu on all those platforms fine | 21:31 |
| tester | @kiko that is correct! | 21:32 |
| kiko | then yes, the tooling supports it | 21:32 |
| kiko | sorry, I was slightly confused by ppcle | 21:32 |
| kiko | as our arch tag is something like ppc64le | 21:32 |
| tester | @kiko how do i sign up? | 21:32 |
| kiko | tester, I just privmsgd | 21:33 |
| prav | I am using MAAS 2.0, but I want to deploy Centos7 to nodes. Is there a doc that I can follow. I am not able to get maas-image-builder | 21:50 |
| prav | I have the Centos7, but the nodes still get the Trusty images | 21:51 |
| kiko | prav, just import centos images | 22:01 |
| kiko | prav, go to the images page and turn centos on | 22:01 |
| prav | I did, but where do I tell MAAS to install Centos instead of Ubuntu? Is it in preseed? | 22:08 |
| prav | Maybe I have got it wrong, but I want to use MAAS to deploy only Centos7 images and using it only for baremetal OS deployments including PXE.. | 22:09 |
| prav | Right now the 3 nodes are taking the IP and getting Ubuntu.. having issues with logging into the nodes, but I think that has got to do with ssh keys not correctly imported.. | 22:10 |
| prav | so when I checked Centos only and unchecked Ubuntu images, the boot failed since it is looking for ubuntu- no such image/boot-kernel failes | 22:11 |
| kiko | prav, when you go to deploy, edit the node | 22:18 |
| kiko | and change the OS from Ubuntu to CentOS | 22:18 |
| kiko | interesting | 22:18 |
| kiko | prav, that's an actual bug | 22:18 |
| kiko | prav, that last piece -- could you file it for us at launchpad.net/maas/+filebug | 22:18 |
| roaksoax | prav: you need ubuntu to deploy centos | 23:09 |
| roaksoax | prav: maas 2.0 will automatically import ubuntu xenial | 23:09 |
| roaksoax | prav: and you can go to the images page | 23:09 |
| roaksoax | prav: and import centos | 23:10 |
| roaksoax | prav: from there you will be able to deploy | 23:10 |
| prav | kiko, I will file that bug | 23:13 |
| roaksoax | prav: 18:20 < prav> so when I checked Centos only and unchecked Ubuntu images, the boot failed since it is looking for ubuntu- no such image/boot-kernel failes | 23:14 |
| roaksoax | prav: that is not a bug | 23:14 |
| roaksoax | prav: to deploy centos we need ubuntu | 23:14 |
| prav | ok. Got it | 23:15 |
| prav | Let me try few options right now | 23:15 |
| roaksoax | prav: so you/win 5 | 23:27 |
| roaksoax | err | 23:27 |
| prav | Is there a doc I can look at step by step for acquiring nodes and deploying? My nodes are now in "Failed commissioning" | 23:36 |
| prav | Loading centos/amd64/generic/centos70/no-such-image/boot-kernel... failed: No such file or directory | 23:39 |
| prav | boot: | 23:39 |
| prav | also complains of power state of manual | 23:44 |
| prav | in maas.log | 23:44 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!