| mup | Bug #1625910 opened: [2.1] MAAS doesn't have routing to BMC IP, but MAAS shows the power as ON <MAAS:New> <https://launchpad.net/bugs/1625910> | 04:22 |
|---|---|---|
| === menn0 is now known as menn0-afk | ||
| sujeet_ | Hi roaksoax | 05:06 |
| sujeet_ | Hi kiko | 05:06 |
| === frankban|afk is now known as frankban | ||
| === menn0-afk is now known as menn0 | ||
| === jamespag` is now known as jamespage | ||
| NewJorg | Hello, is there a way to access the tags of a machine in commissioning? We have some nodes with hardware raids and want to configure the disks differently (RAID0 for every disk or RAID10 for all disks) depending on a tag. | 12:02 |
| rock__ | Hi. MAAS cloud juju bootstrap failed. Followed https://jujucharms.com/docs/2.0/clouds-maas I used MAAS 2.0 on Xenial(16.04). I am using all physical servers. I want to have [MAAS+Openstack base bundle] setup with our own bundle. But we are facing the issue while JUJU bootstrapping the MAAS cloud. | 12:37 |
| rock__ | Complete issue details: http://paste.openstack.org/show/582397/. Please help me in resolving this. | 12:37 |
| rock__ | roaksoax: Can you please help me in this. | 12:37 |
| kiko | lesseee | 12:39 |
| kiko | NewJorg, it should be possible, yes. what does the machine query return to you? | 12:40 |
| kiko | rock__, first, can you deploy Ubuntu manually to all of the nodes in your MAAS? | 12:40 |
| rock__ | kiko: Yes. I deployed and tried. working fine. | 12:41 |
| baldpope | kiko, question on deploying ubuntu via maas - is it still iscsi root? | 12:45 |
| baldpope | so I could use all local disks for something other than / | 12:46 |
| rock__ | kiko/roaksoax: So what I need to do here to resolve the issue. Am I need to change "curtin_userdata" script by putting some sleep before updating apt-get to reflect the change what we made through SSH into that bootstrapping node? | 12:50 |
| kiko | roaksoax, hmm, what change you made through SSH? | 12:50 |
| kiko | baldpope, it never was iscsi root, except for the ephemeral environment. does that make sense or should I explain? | 12:51 |
| roaksoax | rock/win 4 | 12:51 |
| roaksoax | err | 12:51 |
| roaksoax | rock__: as per kiko above, what change you made with SSH | 12:52 |
| rock__ | kiko: To avoid that Hash sum mismatch issue , http://paste.openstack.org/show/582398/ | 12:52 |
| rock__ | roaksoax: I did the above change | 12:52 |
| kiko | rock__, that change doesn't make sense to me | 12:53 |
| kiko | rock__, why did you make it? | 12:53 |
| baldpope | kiko, no, i got it - ty | 12:53 |
| kiko | baldpope, HOWEVER, the caringo team has a set of patches proposed to us that allow you to do exactly that | 12:53 |
| kiko | baldpope, i.e. deploy an OS that is a kernel and initrd only, and then use the local disks for object storage | 12:54 |
| rock__ | kiko/roaksoax: why i added info : http://paste.openstack.org/show/582400/. | 12:56 |
| kiko | rock__, that error points to a wider problem | 12:57 |
| kiko | rock__, is your local mirror not functioning properly? | 12:57 |
| kiko | rock__, or do you have a proxy issue? | 12:57 |
| NewJorg | kiko, what do you mean exactly with machine query? | 12:59 |
| baldpope | kiko, please accept those changes ;) | 13:00 |
| baldpope | that would be exactly what we'd want - use all the local disks for the object storage | 13:01 |
| rock__ | kiko/roaksoax: we got the same hash sum mismatch issue in our local systems , then we genarally used to add that. Local mirror working fine now. | 13:01 |
| samek_ | hi | 13:01 |
| samek_ | how can I debug adding a chassis ? | 13:02 |
| samek_ | we have a moonshot chassis and I want to add it | 13:02 |
| samek_ | All I get is Unable to find a rack controller with access to chassis X.X.X.X | 13:02 |
| rock__ | kiko/roaksoax: No . I don't have any proxy issue. | 13:03 |
| rock__ | kiko/roaksoax: I don't know what to do now. please tell me if you any ideas to resolve this issue. | 13:05 |
| kiko | samek_, there is no rack controller that is active on that subnet | 13:08 |
| kiko | rock__, your problem is not inside MAAS. it is outside MAAS. | 13:08 |
| kiko | rock__, that apt-get update needs to pass without error. your network, proxy or mirror are broken. | 13:09 |
| kiko | rock__, plug a laptop in. does the apt-get upgrade show the same error? | 13:09 |
| samek_ | kiko | 13:11 |
| baldpope | rock__, did you define a proxy in maas? | 13:11 |
| baldpope | if so, does your network allow the internal network to talk with the proxy? | 13:12 |
| rock__ | kiko: sorry. Didn't get. Please tell me clearly. | 13:12 |
| rock__ | baldpope: Didn't define proxy. | 13:12 |
| samba35 | can some one please give me link to read how to /start using kind of stuff for maas ,i am on ubuntu 16.04.1 with maas 2 | 13:13 |
| baldpope | so in that case (I think - kiko, please chime in) the internal hosts are going to attempt direct network access - does the internal network have access to the public internet through your firewall? | 13:13 |
| samek_ | kiko It is accessible from the server where everything is on. | 13:14 |
| rock__ | baldpope: Yes. | 13:14 |
| baldpope | hm | 13:14 |
| baldpope | apologies - I didn't see the paste at openstack - network does not appear to be the issue, your getting out and package info back | 13:16 |
| kiko | baldpope, it is either a network issue, or bad RAM/NIC | 13:17 |
| kiko | rock__, can a laptop, plugged into the same network, using the same mirror, do an apt-get update successfully? | 13:17 |
| baldpope | kiko, line 6/7 - wouldn't that imply a bad file received? | 13:17 |
| rock__ | kiko: $ sudo apt-get upgrade && sudo apt-get dist-upgrade working fine. | 13:18 |
| rock__ | kiko: apt-get update also working fine. | 13:18 |
| baldpope | rock__, but you need to prep by running update - that tells upgrade what to do | 13:18 |
| baldpope | and based on your paste - update is failing | 13:19 |
| rock__ | kiko/baldpope: From MAAS conigured node , I am able to do apt-get update successfully. | 13:20 |
| baldpope | i'm confused - where is apt-get update failing then? | 13:22 |
| rock__ | baldpope: On MAAS slave nodes. | 13:24 |
| baldpope | am I missing something or do your last 2 statements contradict each other? | 13:26 |
| kiko | they do | 13:26 |
| kiko | thanks baldpope :) | 13:26 |
| rock__ | koko/roaksoax/baldpope: when juju bootstrapping the MAAS cloud on one of MAAS Slave node I am getting apt-get issue. | 13:35 |
| kiko | rock__, oh. you are saying that only during a juju bootstrap does the error happen | 13:38 |
| kiko | hmmmmmm | 13:38 |
| kiko | rock__, is the clock on these machines synchronized? | 13:38 |
| rock__ | kiko: please look at this once. http://paste.openstack.org/show/582407/. | 13:43 |
| rock__ | kiko: I think machines are synchronized with the above result. | 13:45 |
| rock__ | Kiko/roaksoax: Exactly I found the issue. When we did juju bootstrap of MAAS cloud on one of MAAS slave node, | 13:47 |
| rock__ | kiko/roaksoax: I ssh into that bootstrapping node and run apt-get update . Then it was giving Hash sum mismatch issue. But after adding http://paste.openstack.org/show/582398/ in that maas slave node apt-get update going successfully. | 13:49 |
| rock__ | kiko/roaksoax: But While it was doing juju bootstrap, we added change is not reflecting immediately. Because curtin script is going very fast. without taking our changes it was giving issues. | 13:51 |
| rock__ | kiko/roaksoax: So can we modify curtin script to put delay before running apt-get update? | 13:52 |
| rock__ | kiko/roaksoax: Please tell me how to edit curtin script if we can. | 13:53 |
| baldpope | rock__, why do you think you need to edit curtain? | 13:58 |
| rock__ | baldpope: whie running curtin script only right we got that issue. | 13:59 |
| baldpope | but (as I understand it) curtain is just setting up your environment, and arguably you'd want to run an apt-get update as one of the first things, right? | 14:00 |
| baldpope | so instead of skipping that step, lets figure out why it's failing | 14:00 |
| rock__ | baldpope: Just look at /etc/maas/preseeds/curtin_userdata | 14:01 |
| rock__ | baldpope: Curtin itself doing apt-get update . there it is failing | 14:02 |
| rock__ | baldpope: Curtin installation failed due to that apt-get update failure | 14:03 |
| neith | anyone knows if I can deploy heat with autopilot or other penstac modules? | 14:06 |
| neith | *Openstack | 14:06 |
| rock__ | kiko/roaksoax: Please give me the reply . | 14:07 |
| kiko | rock__, I have already told you what I can | 14:07 |
| kiko | either it's bad at source (my #1 bet), goes bad over the wire, or goes bad upon reception. | 14:08 |
| rock__ | kiko: OK. Thank you for your support. | 14:22 |
| rock__ | roaksoax: How can I customize the curtin script? | 14:28 |
| neith | kiko: Can I deploy the heat component with juju on an openstack autopilot deployment? | 14:37 |
| rock__ | roaksoax: cloud-init-output.log for my issue: http://paste.openstack.org/show/582417/. If you have any idea please help me. | 14:42 |
| roaksoax | rock__: a problem with your mirror http://ubuntu.cs.utah.edu/ubuntu | 14:58 |
| roaksoax | rock__: maybe your mirror has older packages than the ones in the Ubuntu image | 14:59 |
| roaksoax | or similar | 14:59 |
| lucas__ | good day guys | 15:18 |
| lucas__ | anyone | 15:18 |
| kiko | roaksoax, I've tried to tell rock__ that a few times, but was unable to communicate effectively :-/ | 15:23 |
| kiko | neith, you can't. I think we have juju charms for heat, though, and you can probably deploy manually -- it's not that much harder than using autopilot | 15:24 |
| roaksoax | wom 12 | 15:50 |
| rock__ | roaksoax: Thank you . Actually I tried with different mirrors . But I faced the same issue. OK. I will try to redeploy once again. Thanks . | 16:02 |
| rock__ | kiko: No problem, You provide good info. Thank you. | 16:04 |
| === CyberJacob is now known as zz_CyberJacob | ||
| kiko | rock__, could your RAM or CPU or NIC be bad? | 16:40 |
| === frankban is now known as frankban|afk | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!