| mpontillo | shadoxx: I would check that the IP address is correct and reachable, and that tgtd is up and running on the controller (sometimes it can get stuck; for MAAS 2.3 we're moving away from iSCSI) | 00:36 |
|---|---|---|
| shadoxx | mpontillo: yeah, have restarted tgtd a few times with sysctl. oddly enough, figured out that i can get it to go if i tcpdump the traffic on the controller | 00:36 |
| shadoxx | I think these machines have something weird with interrupts happening | 00:37 |
| shadoxx | They get frozen sometimes up until I start randomly typing on the keyboard. I'm sure it's a quirk with the hardware I'm using (HP), and that there's a kernel parameter, but at this point I just want to get them provisioned | 00:37 |
| shadoxx | This is the third week trying to get this cluster going | 00:37 |
| shadoxx | About to just say screw it and go full OpenStack. | 00:38 |
| shadoxx | also, not sysctl, systemctl | 00:38 |
| mpontillo | shadoxx: you might check https://certification.ubuntu.com/server/models/?vendors=HPE to see if Canonical has certified the hardware, and if so with what firmware versions, etc | 00:38 |
| shadoxx | Well, I just did full firmware+bios updates. They were running software from 2013. | 00:39 |
| mpontillo | heh | 00:39 |
| shadoxx | Getting that ISO was difficult. But I got it. | 00:39 |
| shadoxx | Lol, the Gen8 isn't certified, but the Gen7 is. | 00:40 |
| shadoxx | I'm logged into the first one I tried to provision since the firmware and bios updates | 00:41 |
| shadoxx | It looks promising. | 00:41 |
| mpontillo | shadoxx: great! also, if you think the issue is with iSCSI, would you mind giving the latest MAAS 2.3 beta a try? in that case, we'd bypass iSCSI and send the image directly to the booting node via HTTP | 00:42 |
| mpontillo | (it sounds like it's not iSCSI at this point, probably the firmware...) | 00:43 |
| shadoxx | No, I think it's iscsi. How do I get the beta? Adding the beta repo? | 00:46 |
| shadoxx | Mind you, I'm running my own apt mirror because my MaaS cluster is completely segregated | 00:46 |
| shadoxx | I have one host that has WAN access and is running full repo mirrors for the MaaS Images, Centos 7, and Ubuntu 16.04 | 00:47 |
| mpontillo | shadoxx: what software are you using for apt mirroring? if it's not a full mirror via something like rsync, you may be missing the bootloaders | 00:48 |
| shadoxx | mpontillo: apt-mirror. i was able to bootstrap all the other machines from it | 00:48 |
| shadoxx | I've ran full mirrors before, pretty sure that's not the issue. I'd be willing to check though if you'd coach me on what to check. :D | 00:49 |
| shadoxx | If it's something as simple as missing bootloaders, I'd be in dept | 00:49 |
| shadoxx | debt* | 00:50 |
| shadoxx | Right now the cluster is in PoC mode. | 00:50 |
| shadoxx | So I can run beta if that'll probably fix it | 00:50 |
| mpontillo | shadoxx: if you've installed MAAS via apt and the PPA servers are reachable, you would do something like "apt-add-repository ppa:maas/next" and then do the normal "apt-get update && apt-get dist-upgrade -yu" | 00:51 |
| mpontillo | shadoxx: let me check on the mirror requirements | 00:51 |
| shadoxx | One issue that I *did* run into was that I wasn't pulling down the source repos with my mirror, which was causing all sorts of issue | 00:52 |
| shadoxx | Once I pulled in source repos as well, I was able to commission just fine. | 00:52 |
| mpontillo | shadoxx: I have a post-mirror script that does something like this on my partial mirror http://paste.ubuntu.com/25876414/ | 00:53 |
| mpontillo | shadoxx: basically things broke unless I also grabbed the uefi/ directory for my mirror | 00:54 |
| shadoxx | Let me check if I have that | 00:54 |
| shadoxx | I did not have that. Added 'uefi' to the end of my sources list | 00:58 |
| shadoxx | it's pulling down another 1GB right now. maybe this is it.. | 00:58 |
| mpontillo | shadoxx: ah, I wasn't aware that apt-mirror could treat that like an apt source | 00:59 |
| shadoxx | apt-mirror is pretty great. :] same syntax are your regular sources.list files | 00:59 |
| * mpontillo will keep that in mind for the next partial mirror, thanks | 01:00 | |
| shadoxx | MaaS has been the only thing to make me feel stupid about Linux in the past 2 years. | 01:00 |
| shadoxx | Like, clearly I don't know how this is supposed to work. | 01:01 |
| mpontillo | shadoxx: well, we want to make it as easy as possible =) .. feel free to file bugs if things don't work or aren't intuitive. for example, I feel like we could do a better job raising the issue if bootloaders are missing, rather than finding out when your machines hang at boot time! | 01:04 |
| shadoxx | Well, it's not even that it's hanging. Once a machine is deployed, should it still boot from PXE or should it boot from the main drive? | 01:05 |
| shadoxx | Because, maybe that's why it's failing. I just observed an entire deployment and it seemed fine | 01:05 |
| shadoxx | But it still booted from PXE | 01:06 |
| shadoxx | It looked like there was a step where it checks for a local installation, but I could be misinterpreting the scrolling output | 01:06 |
| mpontillo | shadoxx: that's actually expected, if the machine netboots then MAAS retains some measure of control; the machine might PXE boot and then be instructed to boot local, but if the PXE boot fails the local boot should work fine | 01:07 |
| shadoxx | Gotcha | 01:08 |
| shadoxx | Also, hold off on adding uefi to the mirror.list. I'm figuring out the proper syntax. | 01:08 |
| shadoxx | Should still be able to get it going though... | 01:08 |
| mpontillo | shadoxx: the shell snippet I sent rsyncs the uefi/ directory from the mirror, and then uses "cp -l" to make hard links to it within the pristine mirror directory, which works for my testbed | 01:09 |
| mpontillo | shadoxx: I find it easier to just rsync the entire thing if I have a server with enough space, but I like to take a partial mirror with me on a laptop, because I'm crazy like that ;-) | 01:10 |
| shadoxx | mpontillo: yeah, i found a post that reflects what you're saying too | 01:11 |
| shadoxx | rsync seems to be the only way to go | 01:11 |
| shadoxx | I'll fix this in the morning. I'm about to head out of work and I'm already fatigued with this. | 01:11 |
| shadoxx | But, good news is that UEFI missing might be the ticket. These are Gen8 HP servers. Pretty sure they're UEFI. | 01:11 |
| shadoxx | Would explain why they boot PXE but always fail to boot from local, even after an attempted install. | 01:11 |
| shadoxx | mpontillo: actually, i just modified it to work with my setup pretty easily. I'm going to throw this in a post mirror script | 01:21 |
| shadoxx | boom. done. integrates with my apt -mirror solution nicely. danke | 01:27 |
| shadoxx | testing everything out tomorrow. afk for now. thanks again for the help! | 01:28 |
| mpontillo | shadoxx: let us know how it goes! | 01:45 |
| mup | Bug #1729776 opened: When Maas changes DNS entries, rndc reload fails due to duplicate entries <canonical-bootstack> <MAAS:New> <https://launchpad.net/bugs/1729776> | 06:07 |
| shrekmaxi | Greetings,every one | 06:25 |
| shrekmaxi | Can anyone tell me if MAAS does support debian image import? thanks | 06:33 |
| mup | Bug #1729840 opened: [2.3b3, UI] The SSH keys table in Account is inconsistent with the pattern <2.3qa> <ui> <MAAS:New> <https://launchpad.net/bugs/1729840> | 12:05 |
| mup | Bug #1729841 opened: [2.3b3] I added a SSH key, but it didn't appear in the list until I clicked somewhere on the screen <2.3qa> <ui> <MAAS:New> <https://launchpad.net/bugs/1729841> | 12:23 |
| mup | Bug #1729844 opened: [2.3b3, UI] When I add an SSH key the row should be automatically expanded so that the user can see the key <2.3qa> <ui> <MAAS:New> <https://launchpad.net/bugs/1729844> | 12:23 |
| mup | Bug #1729857 opened: [2.3, UI] Whitespace after checkbox on node listing page <MAAS:Triaged by ack> <https://launchpad.net/bugs/1729857> | 13:56 |
| mup | Bug #1729857 changed: [2.3, UI] Whitespace after checkbox on node listing page <MAAS:Triaged by ack> <https://launchpad.net/bugs/1729857> | 14:02 |
| [Kid] | is there a way to clone nodes in maas? | 14:08 |
| mup | Bug #1729857 opened: [2.3, UI] Whitespace after checkbox on node listing page <MAAS:Triaged by ack> <https://launchpad.net/bugs/1729857> | 14:14 |
| roaksoax | [Kid]: nope | 14:36 |
| mup | Bug #1729868 opened: secondary maas-regiond django errors in log <MAAS:New> <https://launchpad.net/bugs/1729868> | 14:44 |
| mup | Bug #1729694 changed: [2.3, UI, regression] With new node details page, no longer able to see devices (lxc's, kvms) that belong to that machine (e.g. juju cases) <MAAS:Invalid> <https://launchpad.net/bugs/1729694> | 15:32 |
| mup | Bug #1729868 changed: secondary maas-regiond django errors in log <MAAS:Invalid> <https://launchpad.net/bugs/1729868> | 16:11 |
| roaksoax | newell:/win 5 | 17:22 |
| catbus | roaksoax: you are not remotely controlling newell's desktop, right? | 17:27 |
| roaksoax | lol no | 17:37 |
| mup | Bug #1729902 opened: [2.3] Incorrect warning on summary page when recommissioning node with testing <MAAS:Triaged by ltrager> <https://launchpad.net/bugs/1729902> | 17:56 |
| mup | Bug #1729904 opened: [2.3, HWTv2] Add warning icon with a tooltip for overriden status in machine details submeny and the header <MAAS:Triaged> <https://launchpad.net/bugs/1729904> | 17:56 |
| shadoxx | mpontillo: Figured out my issue. I had put my nodes in a different zone than default. It was failing to boot with iscsi for some reason, probably because of the zone | 18:02 |
| shadoxx | once i pulled them out of the zone, i was able to deploy my first node. :] | 18:02 |
| shadoxx | Probably my fault and not a bug. I need to comb over the manual and MaaS concepts a bit more. | 18:02 |
| shadoxx | In any case, the uefi stuff you sent me yesterday probably fixes a problem that I haven't ran into yet, so thanks again | 18:03 |
| mpontillo | shadoxx: sure, it could be that your MAAS already downloaded the UEFI bits, if it had access to them before you started using your internal mirror | 18:08 |
| mpontillo | shadoxx: weird that the zone would affect it, I wouldn't expect that. do you mean DNS zone or availability zone? | 18:08 |
| shadoxx | mpontillo: physical zones. i guess that translates to availability zones? | 18:09 |
| mpontillo | shadoxx: yes, same concept | 18:09 |
| shadoxx | Not a huge deal. They're all in the same availability zone right now anyway. | 18:09 |
| mpontillo | shadoxx: AZs are probably the MAAS feature I've tested the least; probably should step up testing there =) | 18:10 |
| mpontillo | shadoxx: if you can narrow the cause of the issue, I'd encourage you to file a bug | 18:10 |
| shadoxx | definitely | 18:10 |
| shadoxx | I'm going to use one of the nodes as my test node, while leave the other three deployed | 18:11 |
| shadoxx | Now I just need to figure out how MaaS manages these nodes when they're setup as hypervisors. Through Juju I assume | 18:11 |
| mup | Bug #1729913 opened: recent update to 1.9 broke dhcpd.conf.template <1.9.5> <maas> <MAAS:Incomplete> <https://launchpad.net/bugs/1729913> | 18:57 |
| mup | Bug #1729913 changed: recent update to 1.9 broke dhcpd.conf.template <1.9.5> <maas> <MAAS:Incomplete> <https://launchpad.net/bugs/1729913> | 19:03 |
| mup | Bug #1729913 opened: recent update to 1.9 broke dhcpd.conf.template <1.9.5> <maas> <MAAS:Incomplete> <https://launchpad.net/bugs/1729913> | 19:09 |
| mup | Bug # changed: 1514045, 1518633, 1527699, 1546235, 1555335, 1581453, 1603156, 1603469, 1603509, 1604375, 1612248, 1615794 | 21:24 |
| mup | Bug #1614387 changed: [2.0rc4] interface name should be version-sorted (ens3 before ens10) in WebUI <MAAS:Invalid> <https://launchpad.net/bugs/1614387> | 21:33 |
| mup | Bug #1614387 opened: [2.0rc4] interface name should be version-sorted (ens3 before ens10) in WebUI <MAAS:Invalid> <https://launchpad.net/bugs/1614387> | 21:42 |
| mup | Bug #1614387 changed: [2.0rc4] interface name should be version-sorted (ens3 before ens10) in WebUI <MAAS:Invalid> <https://launchpad.net/bugs/1614387> | 21:51 |
| mup | Bug # changed: 1467532, 1516393, 1521376, 1534241, 1545500, 1557569, 1571460, 1578333, 1578347, 1588469, 1691778 | 22:03 |
| mup | Bug # changed: 1445073, 1445939, 1450450, 1450719, 1454024, 1459898, 1462299, 1469742, 1470930, 1472338, 1477691, 1479839, 1695290 | 22:12 |
| mup | Bug # changed: 1427573, 1442234, 1442446, 1442575, 1443257, 1444992, 1447709, 1482405, 1490847, 1491742, 1497991, 1499450, 1504863, 1528628, 1529634, 1531493, 1531531, 1531600, 1536320, 1539292, 1541030, 1541298, 1544308, 1548394, 1549174, 1549331, 1549670, 1552882, 1552892, 1555736, 1556158, | 22:42 |
| mup | 1556963, 1557685, 1559341, 1571101, 1571769, 1619488, 1632638, 1661440, 1664563, 1664618 | 22:42 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!