=== blahdeblah_ is now known as blahdeblah === CyberJacob is now known as zz_CyberJacob [11:47] smoser: that dns v4 first trick fixed it [11:47] thank you :) [13:40] Bofu2U, its fixed without that workaround now. [13:40] ah cool [13:40] you can read the analysis in https://bugs.launchpad.net/ubuntu/+source/squid-deb-proxy/+bug/1547640 [13:41] perfect, much appreciated [13:41] it seems taht squid just kind of goes buggy if a host has > 9 ipv6 addresses, then it never considers ivp4 ones. [13:41] (which it woudl have correctly chosen realizing that it could not connect to any of the ipv6 addresses) [15:09] durschatz: /win 15 [15:09] durschatz: let's do through here :) [15:10] durschatz: do you have any console logs of whent he machine is trying to pxe boot and at the same time logs from /var/log/maas/regiond.log,clusterd.log ? [15:10] yes I do [15:11] what is best way to send them to you? [15:11] durschatz: pastebin [15:13] will do [15:19] var_log_maas_clusterd.logvar_log_maas_regiond.logvar_log_maas_maas.log [15:24] var_log_maas_clusterd.log @ http://pastebin.com/UEiPubyi [15:32] cengnlynxpod1:var_log_maas_regiond.log @ http://pastebin.com/5zUMrUWx [15:32] I had to remove the top and bottom of region.log to fit it into 512k limit [15:34] cengnlynxpod1:var_log_maas_maas.log @ http://pastebin.com/tyYrbgDc [15:37] durschatz: what about clusterd.log and a console output ? [15:43] var_log_maas_clusterd.log @ http://pastebin.com/UEiPubyi [15:44] durschatz: the problem may be this: provisioningserver.drivers.power.PowerError: Failed to power node-87bce854-d7e7-11e5-b6c6-5254009f68e6. BMC never transitioned from to on. [15:45] OK. What could have caused this to start happening after three weeks of no issues. [15:46] B9DE7FD3-5F8E-4786-BE85-773E8F774398-534-0000CE4DEEC657F8 [15:47] for console all I have is a video and an image snapshot of the error thrown by maas-ipmi-autodetect-tool [15:47] do you need these? [15:49] or were you looking for dmesh and/or boot.log [15:49] of which neither has the thrown error. [15:49] dmesg [15:49] durschatz: it would be ideal if you have a console log itself so I can review, but I guess a video of the error thrown would be helpful [15:51] video is 65 MB. how would you like to receive it? [16:01] np ill post it on youtube [16:11] durschatz: yeah or put it somewhere that I can watch it via [16:11] Console Video Ubuntu Mass Fail to Commission @ https://www.youtube.com/watch?v=bOp_HqSjpvg [16:12] durschatz: i know what the issue is [16:13] cool! [16:13] durschatz: the commissioning environment cannot access the internet, and cannot install dependencies it needs to configure the BMC [16:14] OK I will check. [16:16] Is it the node that cannot access the internet? [16:17] Looks like the node can access the internet: [16:17] ubuntu@node1-compute:~$ ping 8.8.8.8 [16:17] PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data. [16:17] 64 bytes from 8.8.8.8: icmp_seq=1 ttl=58 time=7.17 ms [16:18] MAAS also can reach the internet. [16:19] durschatz: oh, I know what this may be about. can you change the power parameters manually on your system, and can you re-try commissioning? [16:19] durschatz: if you've seen just 3/4 days ago, then this is related an issue with the archives/squid that we are tracking down now [16:19] durschatz: we as in ubuntu, provided it is not a maas issue exactly [16:20] that sounds like more of the problem I was expecting. [16:20] When you say change power parameters on my system manually. What should I change them to? [16:21] durschatz: do you have the default power params for the BMC ? [16:21] durschatz: the defuatl user/password combination [16:23] I am using LAN_2_0 [IPMI 2.0] with login and password [16:33] Sorry Andres, I'm not sure what you're asking me to do? For power type, there is not default to select. [17:09] Bug # opened: 1548387, 1548390, 1548391, 1548394, 1548396 [17:39] Bug #1548402 opened: unable to deploy nodes w/ 14.04 and MAAS 1.10 on Xenial [19:01] hi all, does anyone have an ideas why in my virtualbox testing environment for MAAS 1.9.0 (one region, two controllers), the machines on the second controller when using bonds are not getting a default gateway? If I break the bond and use eth0, I get a default gateway, but when using bonds I don't. We have this setup working somewhere else on all physical servers [19:14] Thanks Andres, for letting me know you wanted me to set IPMI username to default ADMIN/ADMIN for Supermicro. [19:15] I recommissioned the SuperMicro node and this worked in the sense that the node is now in ready state. [19:17] What about the KVM node? It's using Virsh, not IPMI. [19:17] durschatz: In the UI, you can "Add Chassis" [19:25] roaksoax: I have that setup and it FAILED_COMMISSIONING on Friday. Going to quickly re-try to make sure the problem still exists. [20:11] roaksoax, is it expected that a maas controller is somehow present on all subnets that I want to deploy machines to? I fear that's a far-fetched requirement, if so. If not, then great, and that means I should just be able to go ahead and add subnets using the CLI, right? === zz_CyberJacob is now known as CyberJacob === CyberJacob is now known as Guest65046 [20:25] [ERROR] Failed to upload leases: 'str' object has no attribute 'mac' < when deploying via Landscape. any hints? [20:48] I guess I just have to try this... [20:48] or does anybody have some experience with it? [20:56] nagyz: I think a MAAS controller has to be on any subnet you want PXE on [20:57] I second koaps: MAAS needs to be on subnet that the node is looking to PXE on. [20:58] right, but it's not about PXE [20:58] durschatz: you wouldn't have any ideas on my issue would you? [20:59] what if I'd like to define different subnets (10.1.1.0/24, VLAN 123), and then PXEboot from something else (172.16.1.0/24, VLAN 333), then set up bond0.123 to a static IP? [20:59] surely the controller doesn't have to have an IP on a subnet where we're not PXEbooting from [20:59] the boot requirement is obvious :) [21:01] I'm not 100% sure. PXE happens at layer 2 [21:01] nagyz: I do that now, bond0 is native vlan where PXE is and I add in two more bond0.123 bond0.321 fabrics, in that case MAAS has no idea about the subnets on bond0.123 or bond0.321, it just adds the interfaces, you can create subnets for those vlans using the maas commands [21:02] but MAAS controller doesn't have interfaces on them [21:03] ok, great, exactly the info I'm looking for [21:03] so the fabric in 1.9+ gives me the L2 domain, basically? [21:03] then I need to tie the subnet to a fabric? [21:03] ya [21:04] great, thanks [21:04] anything particular to look out for when creating the subnets from the CLI? [21:04] give me a sec and I can give you the maas commands [21:26] koaps, ?:P [21:26] ah === blahdeblah_ is now known as blahdeblah [23:07] Bug #1548542 opened: maas.drivers.power.ipmi: [WARNING] Failed to change the boot order to PXE ipmi_ctx_open_outofband_2_0: BMC busy on Dell 12G servers [23:58] Bug #1548560 opened: twisted 16 removes deprecated attributes in Request and DummyRequest