/srv/irclogs.ubuntu.com/2017/04/05/#maas.txt

mupBug #1679882 opened: [2.1] rackd status is stuck at "NN% connected to region controllers" with Active-Active regiond <MAAS:New> <https://launchpad.net/bugs/1679882>02:27
mupBug #1662404 changed: "No rack controllers can access the BMC of node: <node name>" when using OpenStack nova power driver  <MAAS:Fix Released> <MAAS 2.1:Triaged> <https://launchpad.net/bugs/1662404>02:45
mupBug #1679887 opened: [2.1] regiond should not rely on maas-rack command <MAAS:New> <https://launchpad.net/bugs/1679887>02:45
mupBug #1679887 changed: [2.1] regiond should not rely on maas-rack command <MAAS:New> <https://launchpad.net/bugs/1679887>03:39
hrz_jluHello everybody out there using MaaS,06:57
hrz_jluChicken-egg problem in maas-cli? Or did i miss a point? -- Can not format logical volume06:58
hrz_jluI can create a VG: SYS_VG=$(maas ${PROFILE} volume-groups create ${HOST_ID} name="system-vg" partitions=$(echo ${SYS_VG_PARTITION} | jq '. | .id'))06:58
hrz_jluI can create a LV: ROOT_LV=$(maas ${PROFILE} volume-group create-logical-volume ${HOST_ID} $(echo ${SYS_VG} | jq '. | .id') name="system-root" size=100G)06:58
hrz_jluBut how can i format this LV?06:58
hrz_jlufail #1: creating a partition returns with: {"block_device": ["Cannot create a partition table on a logical volume."]}06:58
hrz_jlufail #2: to format the volume I need system_id, device_id, id. I have system_id and id, but what is the device_id?06:58
hrz_jlu-->   echo ${ROOT_LV} | jq {   "system_id": "eq6fr7",   "size": 99996401664,   "available_size": 99996401664,   "name": "system-vg-system-root",   "resource_uri": "/MAAS/api/2.0/nodes/eq6fr7/blockdevices/9/",   "uuid": "fff3f729-7811-4426-ba48-2ce74427eb58",   "used_size": 0,   "id": 9 }06:58
=== frankban|afk is now known as frankban
zeestrathrz_jlu: Not sure if helpful but, after creating LV's, we read them with "maas PROFILE volume-groups read" to get their id, then format with "maas PROFILE block-device format SYSTEM_ID, LV_ID fstype=ext4"07:39
dakjPmatulis, hi any suggest?08:32
dakjpmatulis, I've used tcpdump but the file is too big, I just saw the line "images-maas-io.sawo.canonical.com.http" but I can't see the error11:28
brendanddakj, you probably want to confirm that the image was entirely downloaded at the time when it freezes11:30
brendanddakj, you can find the downloaded artifacts in /var/lib/maas/boot-resources11:33
jlu-hrzHello everybody out there using MaaS11:34
jlu-hrz2nd try (sorry browser has crashed)11:34
jlu-hrzChicken-egg problem in maas-cli? Or did i miss a point? -- Can not format logical volume11:34
jlu-hrzI can create a VG: SYS_VG=$(maas ${PROFILE} volume-groups create ${HOST_ID} name="system-vg" partitions=$(echo ${SYS_VG_PARTITION} | jq '. | .id'))11:34
jlu-hrzI can create a LV: ROOT_LV=$(maas ${PROFILE} volume-group create-logical-volume ${HOST_ID} $(echo ${SYS_VG} | jq '. | .id') name="system-root" size=100G)11:34
jlu-hrzBut how can i format this LV?11:35
jlu-hrzfail #1: creating a partition returns with: {"block_device": ["Cannot create a partition table on a logical volume."]}11:35
zeestratjlu-hrz: Not sure if helpful but, after creating LV's, we read them with "maas PROFILE volume-groups read" to get their id, then format with "maas PROFILE block-device format SYSTEM_ID, LV_ID fstype=ext411:35
jlu-hrzfail #2: to format the volume I need system_id, device_id, id. I have system_id and id, but what is the device_id?11:35
jlu-hrz-->   echo ${ROOT_LV} | jq {   "system_id": "eq6fr7",   "size": 99996401664,   "available_size": 99996401664,   "name": "system-vg-system-root",   "resource_uri": "/MAAS/api/2.0/nodes/eq6fr7/blockdevices/9/",   "uuid": "fff3f729-7811-4426-ba48-2ce74427eb58",   "used_size": 0,   "id": 9 }11:35
jlu-hrz@zeestrat: thanks I will try it...11:36
zeestratjlu-hrz: Same goes for mounting. "maas PROFILE block-device mount SYSTEM_ID LV_ID mount_point=/your/mount/point"11:37
jlu-hrzmaas ${PROFILE}  volume-groups read ${HOST_ID}11:43
jlu-hrz        "logical_volumes": [             {                 "block_size": 4096,                 "size": 99996401664,                 "available_size": 99996401664,                 "path": "/dev/disk/by-dname/system-vg-system-root",                 "model": null,                 "name": "system-vg-system-root",                 "filesystem": null,                 "used_size": 0,                 "type": "virtual",                 "id_path":11:43
jlu-hrz maas test-ech167 partition format eq6fr7 4 fstype=ext411:43
jlu-hrz--> Not Found11:43
jlu-hrz@zeestrat: doesn't work :-(11:44
jlu-hrzagain (stripped...)11:47
jlu-hrz"logical_volumes": [ { "available_size": 99996401664, "system_id": "eq6fr7", "partition_table_type": null, "tags": [], "used_for": "Unused", "partitions": [], "resource_uri": "/MAAS/api/2.0/nodes/eq6fr7/blockdevices/9/", "id": 9 } ],11:47
jlu-hrz maas test-ech167 partition format eq6fr7 9 fstype=ext411:47
jlu-hrz--> Not Found11:47
jlu-hrz(looks better) but doesn't works as espected...11:48
just_hello to everyone12:01
just_is there any guy who can help me?12:01
=== Sagar is now known as Guest21969
just_hello???12:02
jlu-hrz@just_: we can help you. Please tess us your problem...12:03
jlu-hrztell12:03
Guest21969Hi All, I have Asus hardware with SolarFlare NIC cards. PXE is able to assign IP address via DHCP server, but during deployment ipconfig does not work and hence iscsistart command fails. Any idea how can I configure MAAS to use custom drivers during OS install?12:03
just_Okay12:05
just_I have an installation of maas with version 1.95, we had restarted manually one hipervisor and now cloud-init cant give the ip to the node12:05
just_and its printing a lot of errors like this one12:06
just_handlers.py[WARNING]: failed posting event: start: modules-final/config-power-state-change: running config-power-state-change with frequency once-per-instance12:06
just_we cant reach the hipervisor and we cant connect, if we connect thought kvm i can reach the server12:07
just_but i cant ssh12:07
just_i have ubuntu 1412:07
Guest21969issue observed by me is similar to that reported in https://bugs.launchpad.net/maas/+bug/144211212:08
Guest21969we are running MAAS version 2.112:08
Guest21969any thoughts on what could be the possible issue?12:08
just_i've found this12:09
just_https://bugs.launchpad.net/maas/+bug/166368612:09
just_wich OS do u have?12:10
jlu-hrz@Guest21969: Maybe this ist helpful.. http://stackoverflow.com/questions/35087666/how-to-upload-custom-boot-image-to-maas ?12:11
Guest21969jlu-hrz: I don't want to upload custom boot image, just want to use specific drivers for SolarFlare NICs12:11
Guest21969I am a little confused with the current situation, during commissioning ... network card gets detected and IP address is assigned to the commissioning OS, but drivers are not available during installation of OS12:13
jlu-hrzwithout drivers inside your initramfs you won't be able to start your nic's, right?12:13
just_@jlu-hrz do you have any advice for our problem?12:13
Guest21969I thought in settings there are options available to pass on special drivers for the h/w12:14
just_why we cant install maas 2.2 in ubuntu 14 tho?12:14
jlu-hrzyou are not able to load any driver via network without your NICs .12:14
jlu-hrzright?12:15
Guest21969yes12:15
Guest21969everytime it fails during iscsi lun configuration, so I assume it is because of unable to connect to tgt server running on my MAAS server12:16
jlu-hrzare you able to install (temporary) an other NIC? (eg. USB-NIC)12:17
just_No one can help me?....12:22
Guest21969jlu-hrz: I don't have a USB NIC12:22
Guest21969server is located in the DC12:22
jlu-hrzare you able to connect to the server via SSH(KVM)-console?12:24
just_yes i am12:27
just_i can reach the machine with KVM12:27
just_interfaces are up too...12:27
Guest21969jlu-hrz: yes, I am able to connect to server via console12:29
jlu-hrzzeestrat: I got it: maas test-ech167 block-device format eq6fr7 9 fstype=ext412:31
jlu-hrzthanks12:31
just_:(12:31
jlu-hrzGuest21969: are you able to switch to a rescue-console after error occured?12:35
Guest21969jlu-hrz: it takes me to initramfs prompt12:36
Guest21969is there any quick start on how to build custom MAAS images with custom drivers?12:39
dakjbrendand, the directory is empty12:41
brendanddakj, really? even as maas says it's downloading?12:41
brendanddakj, with ls -a as well?12:42
dakjbrendand, if I try to download 14.04 image the proceed is fine, 16.04 starts that and around 96% it stops.12:43
dakjbrendand, sorry I saw the wrong path,12:43
brendanddakj, ok but i'm trying to find if it's fully downloaded on the region12:43
dakjbrendand, there is that "snapshot-20170404-204225"12:44
dakjBrendan, have a look to my post (http://askubuntu.com/questions/899073/maas-dont-make-the-download-of-ubuntu-16-04-lts-image).12:45
brendanddakj, is the final status what's shown in the second screen?12:47
brendanddakj, i.e. 14.04 is synced and 16.04 is 'selected for download'?12:47
zeestratjlu-hrz: Glad to hear.12:57
Guest21969any thoughts on how to build custom MAAS images?12:58
jlu-hrzGuest21969: Which OS?13:00
dakjBrendan, yes it's13:00
jlu-hrzubuntu: https://ubuntuforums.org/showthread.php?t=141213213:02
dakjBrendand, I've also to re-install all but same issue. I receive that when Canonical has published Openstack Autopilot in 16.04 with conjure-up. Before I never had issue like that13:02
brendanddakj, can you answer my question though?13:03
Guest21969jlu-hrz: thanks for the info. How can we push custom drivers within the image? Do you have any thoughts?13:04
dakjbrendand, sure, yes  the second screen is the final status, and 14.04 is synced the 16.04 no13:05
jlu-hrzGuest21969: maybe this is a good starting point: https://cptyesterday.wordpress.com/2012/06/11/adding-non-free-drivers-to-a-debian-netboot-initrd/13:05
brendanddakj, did you try just restarting the image import?13:06
brendandi.e. click 'Save selection'?13:06
dakjbrendand, yes I did that, and I try also to re-install all but nothing same problem13:06
dakjbrendand, yes, I tried to make two test, first one selected only 16.04, and second one with 14.04. with 14.04 the procedure is completed with 16.04 no13:08
dakjbrendand, if you want I can share my desktop with you via teamviewer!13:09
jlu-hrzGuest21969: have you tried different kernels? hwe etc.?13:11
jlu-hrzAFAIF "hwe" means hardware enhanced13:12
brendandjlu-hrz, no it means hardware enablement13:13
jlu-hrzbrendand: Thanks. do you know if SolarFlare NICs are supported?13:14
Guest21969not tried with all, but can check13:14
jlu-hrzbrendand: Thanks. Do you know if SolarFlare NICs are supported?13:14
brendandjlu-hrz, no i don't, but as you say, trying with a newer kernel may help13:15
Guest21969jlu-hrz: default seems to be hwe-t, what do you recommend?13:17
mupBug #1680086 opened: [2.2, trunk] Discovery of machines in a Pod can have hostnames that colide <rsd> <MAAS:Triaged> <https://launchpad.net/bugs/1680086>13:25
dakjbrendand, now I try to download the 14.04 and it results is in synced status, 16.04 continues to stay in "Selected for download"13:27
cnfhas anyone deployed ESX nodes with MaaS ?13:45
David_Orangehello maas experts, i am facing something anoying on maas2.1.4. I am using MAAS to install a machine on a management network, then manually change its network configuration and change its vlan config in the switch, with a network not manage by maas. All those steps works fine, but at the reboot, the server is looping trying to reach maas with pxe. The only solution i found for now is to switch to management vlan, wait for pxe boot, then switch back to 13:57
cnfDavid_Orange: why would you do that?14:00
mupBug #1667141 opened: [FFe] Standing FFe for MAAS 2.2 <maas (Ubuntu):New> <https://launchpad.net/bugs/1667141>14:01
cnfDavid_Orange: changing the network config, that is14:01
pmatulisdakj, can you elaborate on the connection with autopilot/conjure-up? are you using these things in any way?14:01
cnfDavid_Orange: just set the right config in maas14:01
David_Orangecnf: i want to do that to split networks for security reason. This new server shall also hosts a pxe server (maas or other), so i have to split networks14:03
cnfso why use maas on it, at all?14:03
David_Orangecnf: "on it" = on what ?14:04
cnfjust install it manually14:04
dakjPmatulis, at moment no, but my intention is to try that. if I can't download image of 16.04 I can deploy it to my node14:04
David_Orangecnf: this server is a remote server in a DC, i can not go there each time i need to install a new server14:05
cnfDavid_Orange: that is what IPMI is for14:05
cnfDavid_Orange: you are _already_ manually messing with switch configs and the likes, so it's mostly manually anyway14:06
just_hello to everyone....can someone help me?14:06
just_https://bugs.launchpad.net/maas/+bug/166368614:06
just_im having this problem14:06
just_handlers.py failed posting event cloud-init maas14:06
just_i had restarted one hipervisor and now cloud-init cant give to him their IP14:07
dakjPmatulis, I'm running a my old lab with a host where I installed 16.04 and MAAS, before the published of Autopilot, and the download and the sync of 16.04 is correct.14:07
just_and im unable to login into machine14:07
cnfDavid_Orange: maas kind of expects to be able to control the server, manually moving things around will break it14:07
David_Orangecnf: for the context, i works with narinder gupta on OPNFV, and i need to manage several pods. this is a need. and changing switch config is not something tricky in a SDN context14:08
dakjPmatulis, in that host I don't run the last update14:08
David_Orangecnf: so maas is required at each boot ?14:08
just_no one can help me? :S14:08
cnfDavid_Orange: changing things manually isn't SDN :P14:09
David_Orangecnf: maas is forcing the boot routing to pxe only ?14:09
cnfbut it sounds like you just want a PXE installer14:09
cnfnot Metal as a Service14:09
dakjPmatulis, this one is my old lab (http://pasteboard.co/Er1GuAYT.png) that one is new one (http://pasteboard.co/ErQvNtZ6.png), both are in the same network.14:10
David_Orangecnf: ok, it answers my question, MAAS is required at each start, thanks for your answer14:11
cnfi didn't say that, but i guess you have your answer14:12
narinder#topic MANO and various VNF use cases14:12
David_Orangecnf: so is there any configuration on maas for my need ?14:13
cnfdepends on your need, like i said it sounds like you just want a PXE installer14:13
David_Orangecnf: lets say i just want to have a pxe installer for some nodes, and maas for others, MAAS2 can do it or not ?14:15
cnfyou can hammer in a nail with a wrench, doesn't mean you should14:16
cnfi don't like "lets say"s14:16
cnfbut just change your boot order, and be done with it14:16
cnfbut imo, what you are doing is the wrong way about it14:17
dakjPmatulis, why in the second lab the size of 16.04 is 468.1 MB while first one is 656.6 MB?14:17
cnfDavid_Orange: i would just leave the damn network settings, and put the MaaS / pxe network in port isolation14:20
cnfand stop all manual intervention14:20
David_Orangecnf: if it were only a problem of boot order, the failure of pxe would go to the disk after... and it does not, it loop on requesting pxe to maas14:24
David_Orangecnf: and "put the MaaS / pxe network in port isolation" will requires that the node will have an access to my maas network and my other network, thing that i do not want14:26
cnfit would boot from disk14:26
cnfDavid_Orange: no, it doesn't...14:26
cnfit would have access to your MaaS controller14:27
cnfnothing else14:27
cnfthat's what port isolation is for14:27
David_Orangecnf: i need to bring access to this node without give access to all this network node14:29
cnfyes14:29
cnfthat's what port isolation is for :P14:29
cnfhttps://en.wikipedia.org/wiki/Private_VLAN14:29
David_Orangecnf: thanks, i will see if that can fit our needs14:33
cnfgl14:34
mupBug #1680119 opened: maas should enable netconsole during commissioning <MAAS:Triaged> <https://launchpad.net/bugs/1680119>15:01
dakjPmatulis, third installation on another Host in a different network.....same issue with 16.0415:45
pmatulisdakj, the thing is that if there was a systematic problem of downloading 16.04 images with maas then there would be a large outcry. it seems this is due to a local network problem. i'll ping you when i do my own tests on 2.115:47
mupBug # changed: 1575631, 1633696, 1639839, 1642728, 1661581, 1662948, 1670429, 1670444, 1672718, 1674720, 1674807, 1675095, 1675427, 1675432, 1675468, 1675823, 1675838, 1675844, 1675953, 1676619, 1676921, 1676969, 1676978, 1677029, 1677333, 1677350, 1677573, 1677631, 1677658, 1677713, 1677741,16:16
mup1678038, 1678188, 1678236, 1678298, 1678333, 1679231, 1679247, 1679249, 1679315, 167942716:16
sarah__Hello Im testing MaaS on some virtual machines  with virtualbox and I just couldnt figure out the power type I should be using16:22
tychicuswhen commissioning new nodes I am seeing a few minutes worth of these errors: [  148.510201] cloud-init[2197]: request to http://10.10.10.12/MAAS/metadata//2012-01-01/ failed sleeping 1.: <urlopen error[Errno 111] Connection refused>16:22
sarah__the nodes are all marked ready and when I try to commission them it asks me for a power type16:23
sarah__I tried manual but it doesnt work16:23
dakjPmatulis, so 12.04 is fine 14.04 is fine, 16.04, 16.10 and 17.04 no. Are they any different repo to make the download?16:24
tychicusthe node eventually finishes commissioning successfully, but it add a few minutes to the total commissioning time of each machine16:28
tychicuslooks like it may be a similar issue to https://bugs.launchpad.net/maas/+bug/135430916:30
tychicusalso looks similar to this https://bugs.launchpad.net/maas/+bug/132141716:41
tychicusI running 2.1.3+bzr5573-0ubun16:41
tychicuslooks like the issue was related to how the network cards were configured for the machine in MaaS17:02
tychicusthis was a machine that had been previously commissioned, then marked as broken, then re-commissioned, even though one of the networks was marked as disconnected it was still getting an IP address, disabling DHCP on my "maas-external" fabric resolved the issue17:04
pmatulisdak...17:09
dakjpmatulis, I've resolved that, I don't know why for 14.04 yes and 16.04 no (I've to investigate), our Palo Alto Firewalls stopped that using the "Anti-spyware Protection Profile" identified that as a spyware.....removed the profile the download of 16.04 went fine!!!!17:18
pmatulisdakj, well there you go! a local network problem!17:19
pmatulisdakj, i'm glad you got it resolved17:20
dakjpmatulis, yes but it's strange that a firewall identify that as spyware, before that Canonical publi17:20
dakjPublished that the firewall didn't block that17:21
pmatulisdakj, it would be good if you could get more info on that. are there detailed logs of some kind?17:21
dakjpmatulis, yes it's a promise......1 week ago I could to download via MAAS the 16.04 from this week no!!!!17:22
dakjpmatulis, I've to know why the image has been identified as spyware, and last week no!!!17:23
dakjpmatulis, anyway, thanks a lot for your support I hope to return you with a favour. See you soon with th upgrade17:25
=== frankban is now known as frankban|afk
mupBug #1680175 opened: Traceback during upgrade from maas 2.1.3 to 2.2.0~rc1 <landscape> <MAAS:New> <https://launchpad.net/bugs/1680175>17:53
mupBug #1677382 changed: [2.1.5] Failure to PXE boot Cavium crb-1s Servers - PXE-E99: Unexpected network error. Checkpoint D9 ERROR: Boot option loading failed Checkpoint AD <cdo-qa-blocker> <oil> <MAAS:Invalid> <MAAS 2.1:New> <https://launchpad.net/bugs/1677382>18:20
mupBug #1679206 changed: [2.1.5] Maas fails to deploy hardware-ibm-power8-S822LC system with USB drive plugged in <cdo-qa-blocker> <oil> <curtin:Incomplete> <MAAS:Invalid> <https://launchpad.net/bugs/1679206>18:47
mupBug #1680175 changed: Traceback during upgrade from maas 2.1.3 to 2.2.0~rc1 <landscape> <MAAS:Fix Released> <https://launchpad.net/bugs/1680175>18:47
mupBug #1667141 changed: [FFe] Standing FFe for MAAS 2.2 <maas (Ubuntu):Fix Released> <https://launchpad.net/bugs/1667141>19:25
cmd_pancakesIs it possible to deploy a node with a custom preseed config? or would i have to make a custom image for that20:11
roaksoaxcmd_pancakes: custom preseed, yes20:44
roaksoaxcmd_pancakes: /etc/maas/preseeds/curtin_userdata20:44
roaksoaxwould be for ubuntu20:44
roaksoax/etc/maas/preseeds/curtin_userdata_centos for cento,s etc20:44
cmd_pancakesroaksoax: ok great, thanks! i'll take a look20:51
=== smgoller is now known as Guest74578
tychicusI'm sure I am missing something in the docs, but how do you assign a fabric to particular network interface on a machine using the cli, I would think it is something like "maas $PROFILE interface update system_id id fabric=5"20:52
tychicusthat command will return "Success" but not actually change anything20:54
roaksoaxtychicus: maas <profile> interface update <syustem_id> <nic_id> vlan=<vlan id>20:54
roaksoaxtychicus: 16:54 < tychicus> that command will return "Success" but not actually change anything -> can you file a bug for that ?20:54
tychicusroaksoax: thanks, that did the trick, it looks like as long as the system_id and nic_id are correct you can put in any arguments and it will always return "success"21:00
tychicusI'll get the bug filed21:01
roaksoaxtychicus: thanks!21:03
mupBug #1680266 opened: interface update will return successful with invalid params <maas (Ubuntu):New> <https://launchpad.net/bugs/1680266>22:23
mupBug #1677382 opened: [2.1.5] Failure to PXE boot Cavium crb-1s Servers - PXE-E99: Unexpected network error. Checkpoint D9 ERROR: Boot option loading failed Checkpoint AD <cdo-qa-blocker> <oil> <MAAS:Triaged> <MAAS 2.1:New> <maas-images:New> <https://launchpad.net/bugs/1677382>23:17
mupBug #1680266 opened: interface update will return successful with invalid params <MAAS:New> <maas (Ubuntu):New> <https://launchpad.net/bugs/1680266>23:17
mupBug #1680277 opened: [2.2rc1]  rackd — 75% connected to region controllers. <MAAS:New> <https://launchpad.net/bugs/1680277>23:47
mupBug #1680278 opened: [2.2rc1] Failed to process message for node <MAAS:New> <https://launchpad.net/bugs/1680278>23:47

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!