mup | Bug #1749851 opened: [2.4, UI] Machine card always shows status is locked <vanilla-transition> <MAAS:New> <https://launchpad.net/bugs/1749851> | 00:29 |
---|---|---|
mup | Bug #1749863 opened: [2.4, UI, vanilla] Events page is broken, doesn't show "view full history" events <vanilla-transition> <MAAS:Triaged> <https://launchpad.net/bugs/1749863> | 01:17 |
mup | Bug #1749867 opened: [2.4, UI, vanilla] Cannot delete a repository <MAAS:New> <https://launchpad.net/bugs/1749867> | 01:26 |
hallyn | roaksoax: ok no i don't understand - looking at https://askubuntu.com/questions/942412/how-do-you-statically-asign-an-ip-to-a-commissioned-machine-in-maas - that only works if a node is alredy commissioned. | 01:48 |
hallyn | so if i need to commission a node which for physical switch reasons must be a specific ip address, how do i go about it? | 01:49 |
hallyn | simple enough with dnsmasq.conf, but how do i do it with maas? | 01:49 |
mup | Bug # opened: 1749870, 1749871, 1749872, 1749874 | 01:50 |
hallyn | huh, maybe i just mark it broken | 01:52 |
hallyn | not "classy", but might work | 01:52 |
roaksoax | hallyn: machines in the commissioning environment DHCP from maas | 01:53 |
roaksoax | so they will always get a random ip | 01:54 |
roaksoax | from the dynamic range | 01:54 |
roaksoax | ,achines you deploy | 01:54 |
roaksoax | dont get a ip from the dynamic range | 01:54 |
hallyn | roaksoax: but that's not sensible is it? if i have a particular machien in a particular spot, and i 'add' it through the interface and add its macaddr, | 01:59 |
hallyn | then it makes perfect sense to let me say "that macaddr must have tha tip address" | 02:00 |
hallyn | and indeed you do require the macaddr before commissioning | 02:00 |
roaksoax | hallyn: when you initially register a machine, while you may have a mac address, we don't know where that machine will dhcp from | 02:07 |
roaksoax | hallyn: it will dhcp from the dynamic range | 02:07 |
roaksoax | becuase it will dhcp from where ever it finds a dhcp server | 02:08 |
roaksoax | if you configure your dhcp server to give that machine an ip statically, siure you can do that | 02:08 |
roaksoax | but that's not a supported path in MAAS | 02:08 |
roaksoax | hallyn: that said, if you want to overwrite that behavior, you can create a DHCP snippet | 02:08 |
roaksoax | Settings -> DHCP snippet | 02:09 |
roaksoax | and you can add your mapped entry against the mac address | 02:09 |
roaksoax | maybe you can also do it via the api, e.g. say that X interface should have Y IP address, but not sure how whether that would work, or how it would interact with a deployment | 02:09 |
roaksoax | hallyn: and ip addresses used during commissioning | 02:10 |
roaksoax | hallyn: are re-usable | 02:10 |
roaksoax | and can be thrown away | 02:10 |
roaksoax | hallyn: maybe this will help you with how maas manages ranges: http://spectrum42.com/posts/ip-ranges-in-maas/ | 02:12 |
roaksoax | hallyn: and oyu can see dhcp snippets here: https://docs.maas.io/2.3/en/installconfig-network-dhcp#dhcp-snippets | 02:12 |
mup | Bug #1749877 opened: [2.4, UI, vanilla] Locked machine changes no longer visible <regression> <vanilla-transition> <MAAS:New> <https://launchpad.net/bugs/1749877> | 02:56 |
=== frankban|afk is now known as frankban | ||
mup | Bug #1749962 opened: [2.4] Cannot allocate memory when DescribePowerTypes happen <performance> <MAAS:Triaged by blake-rouse> <https://launchpad.net/bugs/1749962> | 14:19 |
hallyn | roaksoax: thanks | 15:48 |
mup | Bug #1749979 opened: [2.4] UpdateNodePowerState RPC calls causes "Must be one of the node's interfaces." <performance> <MAAS:Triaged by blake-rouse> <https://launchpad.net/bugs/1749979> | 15:49 |
hallyn | so i can hopefully do a dhcp snippet with a group/host entryto do what i need | 15:52 |
hallyn | so i wonder if i do that do i leave the ip address that i reserve in the snippet for a host in the dhcp range? or not? | 15:54 |
mup | Bug #1749851 changed: [2.4, UI] Machine card always shows status is locked <regression> <vanilla-transition> <MAAS:New> <https://launchpad.net/bugs/1749851> | 16:58 |
=== frankban is now known as frankban|afk | ||
mup | Bug #1750007 opened: [2.4, UI] Script result page needs fixing <vanilla-transition> <MAAS:Triaged by ya-bo-ng> <https://launchpad.net/bugs/1750007> | 17:55 |
roaksoax | 5 of 36,543   | 18:03 |
roaksoax | 5 of 36,543   | 18:03 |
roaksoax | wtf | 18:03 |
mup | Bug #1750015 opened: [2.4, UI] Logs tab is missing <vanilla-transition> <MAAS:Triaged by ya-bo-ng> <https://launchpad.net/bugs/1750015> | 18:22 |
mup | Bug #1750020 opened: [2.4, UI] Custom scripts overlap on the settings page <vanilla-transition> <MAAS:Triaged by ya-bo-ng> <https://launchpad.net/bugs/1750020> | 18:22 |
mup | Bug #1750015 changed: [2.4, UI] Logs tab is missing <vanilla-transition> <MAAS:Triaged by ya-bo-ng> <https://launchpad.net/bugs/1750015> | 18:31 |
mup | Bug #1750020 changed: [2.4, UI] Custom scripts overlap on the settings page <vanilla-transition> <MAAS:Triaged> <https://launchpad.net/bugs/1750020> | 18:32 |
mup | Bug #1750015 opened: [2.4, UI] Logs tab is missing <vanilla-transition> <MAAS:Triaged by ya-bo-ng> <https://launchpad.net/bugs/1750015> | 18:35 |
mup | Bug #1750020 opened: [2.4, UI] Custom scripts overlap on the settings page <vanilla-transition> <MAAS:Triaged> <https://launchpad.net/bugs/1750020> | 18:35 |
blizzow | I'm running maas 2.3.0 on ubuntu 16.04. I'm trying to convert my machines to boot over UEFI. When I set a new server to boot off UEFI, it gets a PXE boot image from my maas server and gets to just a grub prompt. | 19:02 |
blizzow | Anyone here know if I'm missing something major to get a server booting off MAAS with UEFI? | 19:03 |
doug00 | Hello, does anyone have any information on how to pxe boot ppcle machines with MAAS? | 20:12 |
roaksoax | blizzow: are you using secure boot ? | 20:20 |
roaksoax | doug00: by ppcle you bean ppc64el ? | 20:21 |
roaksoax | doug00: and these actual power 8 physical systems ? | 20:21 |
doug00 | yeah | 20:21 |
doug00 | we have some power8 machines i'd like to put into maas | 20:22 |
roaksoax | doug00: well, you should only need to import the ppc64el images and that should be it really | 20:22 |
doug00 | hmmm I have the images in there but when I tried to pxe boot the server it said something about a pxe error | 20:23 |
doug00 | should there be a ppc64el pxe loader on the maas server? | 20:24 |
roaksoax | doug00: console logs would be helpful. maybe an issue with the firmware version? We CI ppc64el and works just fine | 20:24 |
roaksoax | doug00: when you download the image, and the image gets downloaded into the rack | 20:24 |
roaksoax | doug00: it will download everything you need | 20:24 |
roaksoax | including bootloaders for petitboot to work | 20:24 |
doug00 | hmmm ok, maybe i just need to change the pxe binary in the ppc server. Its pointing to pxelinux.0 which looks like a x86 binary | 20:25 |
roaksoax | doug00: petitboot should be booting "bootppc64.bin" | 20:26 |
doug00 | ok cool | 20:26 |
doug00 | i'll try that | 20:26 |
roaksoax | from MAAS' dhcp: | 20:26 |
doug00 | thanks! | 20:26 |
roaksoax | } elsif option arch = 00:0C { | 20:26 |
roaksoax | # open-firmware_ppc64el | 20:26 |
roaksoax | filename "bootppc64.bin"; | 20:26 |
roaksoax | doug00: if it is an LPAR, it would boot pxelinux.0 | 20:27 |
roaksoax | } elsif option arch = 00:0C { | 20:27 |
roaksoax | # open-firmware_ppc64el | 20:27 |
roaksoax | filename "bootppc64.bin"; | 20:27 |
roaksoax | } elsif option arch = 00:0E { | 20:27 |
roaksoax | # powernv | 20:27 |
roaksoax | filename "pxelinux.0"; | 20:27 |
roaksoax | option path-prefix "ppc64el/"; | 20:27 |
roaksoax | sorry for the paste :) | 20:27 |
blizzow | roaksoax: I dont' think so. | 20:27 |
doug00 | haha no worries | 20:28 |
roaksoax | blizzow: AFAIK, there's a bug in grub with secure boot enabled, which would cause it to go into grub's prompt | 20:28 |
roaksoax | blizzow: https://bugs.launchpad.net/maas/+bug/1711203 | 20:29 |
blizzow | hmm. how would I tell if I'm using secureboot? | 20:29 |
roaksoax | blizzow: depends on the firmware really. the only i have access right now are NUC's and they have an option to enable/disable secure boot | 20:29 |
blizzow | roaksoax: My node doesn't even make it to enlisting... | 20:30 |
doug00 | roaksoax: i see powernv in there as well that points to pxelinux.0 | 20:34 |
doug00 | if option arch = 00:0E { # powernv filename "pxelinux.0"; option path-prefix "ppc64el/"; | 20:34 |
doug00 | i'm guessing that is why my servers are getting pxelinux.0 instead of the bootppc64.bin | 20:34 |
doug00 | how do I know what option my server is sending? | 20:37 |
doug00 | yeah, it looks like my machines are sending the arch = 00:0E and getting powernv portion | 20:50 |
doug00 | roaksoax: its interesting, your dhcpd conf is in an different order than mine, did you modify it? | 21:11 |
doug00 | so far nothing has helped with the pxe booting of the power8 server | 21:37 |
doug00 | i keep getting the following error in syslog "Feb 16 16:35:54 maas sh[1484]: #011builtins.AttributeError: 'Port' object has no attribute 'socket'" | 21:38 |
roaksoax | doug00: nope, mine is just maas created one | 21:39 |
roaksoax | doug00: that is not a issue that will affect you | 21:39 |
roaksoax | doug00: that socket thing | 21:39 |
doug00 | hmmm | 21:39 |
roaksoax | doug00: and if your machine is an LPAR inside power8, the dhcp client will send 00:0E in the client architecture code. if it is deploying ontop of the full system, it should send 00:0C | 21:40 |
roaksoax | if it is not doing that | 21:41 |
roaksoax | that would be a firmware issue | 21:41 |
doug00 | i thought i was deploying ontop of the full system but it keeps sending 00:0E | 21:41 |
roaksoax | then it is probably an LPAR | 21:42 |
roaksoax | unless there's a bug in their firmware | 21:42 |
doug00 | the firmware is from Mar 10 2016 so possibly | 21:42 |
roaksoax | wuld have to check which firmware we are using | 21:43 |
blizzow | Is it possible to boot and use RAID1 for the primary and boot partitions with UEFI? | 22:02 |
TJ- | blizzow: Yes | 22:04 |
blizzow | Hrm, when I commission a new machine running UEFI, MAAS automagically creates a 536.9MB fat32 partition mounted at /boot/efi. If I delete that and create two partitions and make a raid1 then format it fat32 and mount it at boot/efi, then create a second 200GB raid1, it just boots to grub and doesn't get past it. | 22:07 |
blizzow | What would the proper way be to set up RAID1 for both boot and OS? | 22:07 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!