| mup | Bug #1511589 opened: maas provider, hwclock out of sync means juju will not work <kanban-cross-team> <landscape> <maas-provider> <juju-core:New> <MAAS:New> <https://launchpad.net/bugs/1511589> | 02:41 |
|---|---|---|
| digitalsky | hi, my maas server is having trouble detected nodes, am i'm not sure how I can debug this | 03:46 |
| digitalsky | I have 5 baremetal servers on Softlayer, they all have 4 interfaces. eth0 / eth2 on each host are on the private network, and eth1 / eth3 are on the public network. there is another interface for IPMI on the private network. | 03:47 |
| digitalsky | if i add the hardware manually from the maas gui with the IPMI port's IP / mac, then i seem to be able to control the node. the node also seem to be able to get dhcp from maas. but i don't understand why the nodes were not auto detected | 03:49 |
| digitalsky | the maas i'm running is MAAS Version 1.8.3+bzr4053-0ubuntu1 (trusty1) | 03:49 |
| roaksoax | digitalsky: the nodes are not autodetect ? | 03:49 |
| digitalsky | exactly | 03:49 |
| roaksoax | digitalsky: or IPMI is not autodetected ? | 03:49 |
| roaksoax | digitalsky: if the nodes are not autodetected, possible issues: 1. The machine never PXE booted from MAAS, which prevented auto-registration to happen | 03:50 |
| digitalsky | um i'm not aware MAAS would detect the ipmi. i started maas, configured the networks and expected to see it detect the nodes (which have no OSes installed) | 03:50 |
| roaksoax | digitalsky: 2. something happen during the autoregistration process that prevented it from registering itself in MAAS, which would require me to look at the console logs | 03:50 |
| roaksoax | digitalsky: right, so this is what should happen | 03:51 |
| roaksoax | digitalsky: 1. turn on manually a node | 03:51 |
| roaksoax | 2. node will dhcp and PXE from MAAS | 03:51 |
| roaksoax | 3. Node will run the "auto enlistment" process | 03:51 |
| roaksoax | 4. Node will register itself in MAAS | 03:51 |
| digitalsky | ah ok! I thought as soon as the node boot up and get dhcp, then it will already show up as detected. | 03:52 |
| roaksoax | digitalsky: so first thing to check would be whether the nodes are actually DHCP'ing from MAAs and PXE booting | 03:52 |
| roaksoax | digitalsky: nope, it will take 5-15 mins (approx) to do so depending on the type of machine you are using | 03:52 |
| digitalsky | ok i used KVM to see what's going on in one of the nodes, and I see that it is getting an IP from DHCP as well as booting from PXE. however it got to a point where it tried to contact 169.254.169.254 and failed after timing out. | 03:53 |
| digitalsky | subsequently, it continues to boot up to a point where there is a login prompt | 03:53 |
| digitalsky | and then on the MAAS side, it says failed comissioning. probably because the node failed to get the user-data from MAAS so it didn't get set up properly and only booted a plain ubuntu and nothing else | 03:53 |
| digitalsky | looks like it's similar to this: http://tinyurl.com/p264fj2 | 03:56 |
| digitalsky | roaksoax: and i don't think there is an /etc/cloud path in my maas host | 04:00 |
| roaksoax | digitalsky: can you tell me what are the kernel params the system gets hwen PXE booting? | 04:06 |
| roaksoax | digitalsky: that means that they cannot contact the MAAS server for whatever reason | 04:06 |
| digitalsky | roaksoax: yea that's what I figured is that they can't get to the maas for the user-data. let me see if i can get the params | 04:07 |
| digitalsky | roaksoax: would i be able to see some dhcp logs on the maas to see what params it sent out? | 04:08 |
| roaksoax | digitalsky: nope, it is either there, or in the kernel while loading the ephemeral image | 04:08 |
| roaksoax | digitalsky: if not, do: sudo dpkg-reconfigure maas-cluster-controller and make sure it is pointing to the correct IP address (which should be in the same network as the machines that are DHCP'ing) | 04:10 |
| roaksoax | digitalsky: and try again | 04:10 |
| digitalsky | roaksoax: ok. | 04:13 |
| roaksoax | digitalsky: need to run now,past midnight for me, but feel free to leave your findings and will read in the morning | 04:17 |
| digitalsky | roaksoax: thanks | 04:18 |
| digitalsky | roaksoax: i'm looking for cloud_init right? it's pointed to the wrong ip | 04:18 |
| digitalsky | roaksoax: so i'll try to reconfigure it with dpkg and try again | 04:18 |
| digitalsky | roaksoax: i know you have to run, thanks very much for your help so far! :) | 04:18 |
| mup | Bug #1511610 opened: TestReleaseAutoIPs.test__calls_update_host_maps_for_next_ip_managed_subnet can fail randomly <unit-tests> <MAAS:Triaged> <https://launchpad.net/bugs/1511610> | 04:29 |
| mup | Bug #1511689 opened: maasserver.tests.test_forms_nodegroup.TestNodeGroupDefineForm.test_rejects_invalid_interface fails spuriously <tests> <MAAS:Triaged> <https://launchpad.net/bugs/1511689> | 10:54 |
| mup | Bug #1511689 changed: maasserver.tests.test_forms_nodegroup.TestNodeGroupDefineForm.test_rejects_invalid_interface fails spuriously <tests> <MAAS:Triaged> <https://launchpad.net/bugs/1511689> | 11:09 |
| mup | Bug #1511689 opened: maasserver.tests.test_forms_nodegroup.TestNodeGroupDefineForm.test_rejects_invalid_interface fails spuriously <tests> <MAAS:Triaged> <https://launchpad.net/bugs/1511689> | 11:15 |
| mup | Bug #1511690 opened: maasserver.websockets.tests.test_listener.TestNodeFilesystemgroupListener.test__calls_handler_with_update_on_create fails spuriously <tests> <MAAS:Triaged> <https://launchpad.net/bugs/1511690> | 11:24 |
| mup | Bug #1511713 opened: udev rules not updated to reflect MAC change in node <MAAS:New> <https://launchpad.net/bugs/1511713> | 12:24 |
| mup | Bug #1511794 opened: node-group probe-and-enlist-hardware model=vmware does not return errors <kanban-cross-team> <landscape> <MAAS:New> <https://launchpad.net/bugs/1511794> | 17:22 |
| mup | Bug #1511794 changed: node-group probe-and-enlist-hardware model=vmware does not return errors <kanban-cross-team> <landscape> <MAAS:New> <https://launchpad.net/bugs/1511794> | 17:25 |
| mup | Bug #1511794 opened: node-group probe-and-enlist-hardware model=vmware does not return errors <kanban-cross-team> <landscape> <MAAS:New> <https://launchpad.net/bugs/1511794> | 17:28 |
| mup | Bug #1511493 opened: Should not allow partitions to be created on bcache device <curtin:Invalid> <MAAS:Triaged> <https://launchpad.net/bugs/1511493> | 18:28 |
| mup | Bug #1511860 opened: IPMI credentials not set in node power <falkor:New> <MAAS:New> <https://launchpad.net/bugs/1511860> | 21:16 |
| mup | Bug #1511860 changed: IPMI credentials not set in node power <falkor:New> <MAAS:New> <https://launchpad.net/bugs/1511860> | 21:22 |
| mup | Bug #1511860 opened: IPMI credentials not set in node power <falkor:New> <MAAS:New> <https://launchpad.net/bugs/1511860> | 21:35 |
| digitalsky | roaksoax: Hi roaksoax, I updated the MAAS configuration with the internal IP and now all 4 nodes got detected. | 22:26 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!