alexlist | kiko: That is not what I am looking for... for Ubuntu, you can have custom preseeds. We need something similar for CentOS as well, without going through the pain of creating custom images... | 00:18 |
---|---|---|
roaksoax | alexlist: what are you looking for? | 00:24 |
roaksoax | alexlist: you *can* have custom preseeds for centos | 00:27 |
mup | Bug #1588116 opened: [2.0-bzr5059] Adding a rack using dpkg-reconfigure isn't intuitive <MAAS:New> <https://launchpad.net/bugs/1588116> | 01:47 |
mup | Bug #1588125 opened: Nodes commission, but don't see network interfaces with Xenial <cpe-sa> <MAAS:New> <https://launchpad.net/bugs/1588125> | 02:44 |
mup | Bug #1588154 opened: Deploying node fails when it has VLAN configuration <MAAS:New> <https://launchpad.net/bugs/1588154> | 04:58 |
mup | Bug #1588161 opened: Empty boot-source-selections imports all available images <cdo-qa> <MAAS:New> <https://launchpad.net/bugs/1588161> | 06:01 |
=== frankban|afk is now known as frankban | ||
mup | Bug #1588350 opened: [2.0b5] Multiple power query's in too little time <MAAS:New> <https://launchpad.net/bugs/1588350> | 13:00 |
mup | Bug #1588350 changed: [2.0b5] Multiple power query's in too little time <MAAS:New> <https://launchpad.net/bugs/1588350> | 13:06 |
mup | Bug #1588350 opened: [2.0b5] Multiple power query's in too little time <MAAS:New> <https://launchpad.net/bugs/1588350> | 13:12 |
=== jgrimm-afk is now known as jgrimm | ||
kiko | alexlist, what roaksoax said -- you can generate custom preseeds for centos; the preseeds are not OS-specific but rather specific to curtin (and cloud-init) | 15:28 |
roaksoax | kiko: /etc/maas/preseeds/curtin_userdata_centos | 16:00 |
roaksoax | alexlist: ^^ | 16:00 |
roaksoax | alexlist: so you can basically do what you need via preseeds | 16:00 |
gimmic | in maas, can I set the default ipmi credentials? | 16:05 |
gimmic | somewhere | 16:05 |
gimmic | rather than waiting to assign it to hosts which are discovered | 16:05 |
gimmic | I'd prefer it if Maas defaulted to ipmi 2.0 and used N credentials | 16:05 |
roaksoax | gimmic: you can change the credentials after you commission the machine | 16:40 |
=== frankban is now known as frankban|afk | ||
gimmic | so I've got a host deployed, but I can't seem to ssh to it with the set key | 17:02 |
gimmic | actually, status says "deploying" | 17:02 |
kiko | gimmic, need to wait until it's deployed | 17:07 |
gimmic | I can't see any activity on the host or the maas server | 17:07 |
shewless | Hi guys, I changed the "external" facing IP of my maas server and now I'm having trouble. I can connect via ssh but dhcpd isn't running and I dns look-ups to external sites fail. Any hints at config I'd have to change? | 17:35 |
shewless | maas 2.0 | 17:35 |
shewless | ah..nevermind I got it figured out. thanks | 17:40 |
mup | Bug #1588466 opened: gpg --batch --verify during maas install causes Unhandled Error <cdo-qa> <MAAS:New> <https://launchpad.net/bugs/1588466> | 17:44 |
mup | Bug #1588469 opened: cluster view won't load if a single cluster controller lacks a valid UUID <MAAS:New> <https://launchpad.net/bugs/1588469> | 17:44 |
shewless | hmm.. I spoke to soon. I fixed the dns lookups but dhcpd is not starting | 17:48 |
shewless | I can't seem to find any log files related to maas-dhcpd or dhcp | 17:50 |
shewless | any pointers? | 17:50 |
kiko | shewless, you may need to do a dpkg-reconfigure -- you used to in 1.9. roaksoax, any clue? | 18:16 |
shewless | kiko: thanks - is there harm in my doing that on 2.0.0? | 18:27 |
shewless | kiko: also which package would I reconfigure? | 18:28 |
kiko | shewless, I'd guess maas-rackd | 18:29 |
roaksoax | shewless: have you tried setting upstream DNS server in MAAS ? | 18:31 |
shewless | roaksoax yes I have | 18:33 |
shewless | roaksoax: it's set to an internal DNS that is still reachable by ping | 18:34 |
roaksoax | shewless: so why is dhcp not running ? have you looked at systemd ? | 18:34 |
shewless | roaksoax: I'm not sure how to do that. | 18:34 |
roaksoax | shewless: journalctl -u maas-dhcpd -> try that ? | 18:36 |
shewless | roaksoax: 1 single entry: Jun 02 14:37:22 maas systemd[1]: Stopped MAAS instance of ISC DHCP server for IPv4. | 18:37 |
shewless | if I restart the service it just prints that | 18:37 |
roaksoax | shewless: so what if you do ps faux | grep dhcpd ? | 18:38 |
shewless | nothin | 18:38 |
shewless | roaksoax: that command shows me tthat dhcpd is not running | 18:39 |
roaksoax | shewless: ok, so is DHCP enabled at all in MAAS ? | 18:39 |
shewless | roaksoax: yes it is. I tried disabling it and re-enabling it but it didn't seem to have an effect. Related if I do a "service maas-dhcpd status" I see that /var/lib/maas/dhcpd.conf is missing.. that's probably bad right? | 18:40 |
mup | Bug #1588489 opened: package maas-region-controller 2.0.0~beta3+bzr4941-0ubuntu1 failed to install/upgrade: subprocess installed post-installation script returned | 18:41 |
mup | error exit status 1 <amd64> <apport-package> <need-duplicate-check> <third-party-packages> <xenial> <maas (Ubuntu):New> <https://launchpad.net/bugs/1588489> | 18:41 |
shewless | roaksoax: so any idea why the dhcpd.conf file isn't being generated? Is that rackd that would do that? | 19:33 |
kiko | shewless, it's a bug I think -- roaksoax? | 19:47 |
mup | Bug #1532478 changed: Can't add SM15K chassis - can't input management IP address of SM15k <oil> <MAAS:Invalid by newell-jensen> <https://launchpad.net/bugs/1532478> | 19:50 |
kiko | shewless, can you disable and reenable DHCP management on the vlan you are servicing? | 19:52 |
kiko | have you already> | 19:52 |
kiko | ? | 19:52 |
sjl | trying to troubleshoot maas API with Postman. Anyone able to get this basic GET call to work via Postman? http://<insert-your-maas-server-ip>/MAAS/api/version | 19:57 |
kiko | sjl, I have never tried that -- does Postman do OAuth? | 19:57 |
sjl | yes 1.0 and 2.0 | 19:58 |
gimmic | so when deploying, it installs and ends at "cloud-init finished..." | 19:58 |
gimmic | but the box stays in "deploying" state | 19:58 |
gimmic | closer look appears to be "Node installation - 'curtin' failed: configuring disk: sda" | 19:59 |
kiko | yep | 20:00 |
kiko | why did it fail to configure? | 20:00 |
gimmic | where can I find relevant log data? | 20:01 |
kiko | look through the install log | 20:01 |
roaksoax | shewless: have you enabled DHCP ? | 20:02 |
roaksoax | shewless: MAAS will only generate a dhcpd.conf is DHCP is enabled | 20:02 |
roaksoax | shewless: also, pastbin rackd.log | 20:02 |
gimmic | wheres the install log located? | 20:03 |
gimmic | I don't have access to the iron(yet), and don't see anything specific in maas' log | 20:03 |
kiko | gimmic, maas sends it back and it gets attached to the node page IIRC | 20:05 |
gimmic | the node page was where I pulled that from, I don't see any additional detail | 20:05 |
kiko | hm | 20:05 |
kiko | gimmic, are you doing anything custom to sda? | 20:05 |
kiko | and if not, is there anything unusual about it? | 20:06 |
gimmic | generic out-of-box install | 20:06 |
gimmic | r620, single SATA | 20:06 |
gimmic | this will be the second node deployed, the first onedid the same thing until I rebooted it | 20:07 |
gimmic | I'm obviously trying to streamline it a bit and tshoot things | 20:07 |
gimmic | but basic 14.04 LTS image, very generic hardware. I have 30 of these nodes to do | 20:07 |
kiko | gimmic, it's very odd tbh | 20:15 |
kiko | gimmic, so the entire install log has no hints other than the sda failure? can you put it in a pastebin | 20:15 |
gimmic | anywhere I can view the log in actual raw format | 20:16 |
gimmic | rather than the weberface | 20:16 |
gimmic | doesnt seem to be much useful in /var/log/maas | 20:16 |
gimmic | one other interesting question is that on some nodes I want to use a software raid to mirror two disks | 20:18 |
gimmic | but since it uses one off the bat, I can't seem to use maas to configure it to mirror the disk | 20:18 |
kiko | gimmic, you can just configure the software raid in the MAAS UI itself | 20:20 |
kiko | that should be pretty straightforward (and worked as far back as 1.9x) | 20:20 |
gimmic | but not before the node has been commissioned as far as I can tell | 20:20 |
kiko | correct | 20:21 |
kiko | the node needs to commission first | 20:21 |
kiko | oh | 20:21 |
kiko | are you stuck on the commissioning step? | 20:21 |
kiko | commissioning has an option you can select to let you log in and inspect | 20:21 |
kiko | have you checked that? | 20:21 |
kiko | (sorry, I thought you were mid-deployment) | 20:21 |
gimmic | oh, found it. I'm conflagulating two different 'issues' right now. | 20:23 |
gimmic | One set of nodes has two disks, and they will be mirrored / | 20:23 |
gimmic | that I just found under a node that was in ready state | 20:23 |
gimmic | Can I make templates for this somehow? | 20:23 |
kiko | not yet | 20:29 |
kiko | gimmic, you will be able to, but for the moment the solution is to script node config via the API | 20:29 |
gimmic | this is all the detail I can find on the deploy errors: | 20:35 |
gimmic | Node installation - 'curtin' failed: configuring storageThu, 02 Jun. 2016 15:23:01 | 20:35 |
gimmic | Node installation - 'curtin' failed: configuring disk: sdaThu, 02 Jun. 2016 15:23:01 | 20:35 |
gimmic | Node installation - 'curtin' started: configuring disk: sdaThu, 02 Jun. 2016 15:23:00 | 20:35 |
gimmic | Node installation - 'curtin' started: configuring storageThu, 02 Jun. 2016 15:22:59 | 20:35 |
shewless | roaksoax: I'll workon getting rackd log. DHCP is enabled as far as the UI tells me.. on my PXE network | 20:35 |
gimmic | kiko: oddly, if I reboot the node it deploys fine the second time | 20:38 |
gimmic | so it is something with partition creation or an error in the install/deploy script | 20:38 |
gimmic | about to test the theory with a third node | 20:39 |
kiko | wwird | 20:39 |
mup | Bug #1519358 changed: Non-fatal error message observed in PXE boot sequence <MAAS:New> <https://launchpad.net/bugs/1519358> | 20:42 |
mup | Bug #1588531 opened: [2.0b5] Deployed regions should be able to transistion back to machines <MAAS:In Progress by ltrager> <https://launchpad.net/bugs/1588531> | 20:42 |
shewless | roaksoax: here is rackd.log: http://paste.ubuntu.com/16930706. Note: for some reason if resolve.conf points to maas' own DNS server I can't resolve external DNS (like google). DOn't know if that is related | 20:42 |
gimmic | Jun 2 20:47:59 ES18 [CLOUDINIT] util.py[DEBUG]: Running scripts-user (<module 'cloudinit.config.cc_scripts_user' from '/usr/lib/python2.7/dist-packages/cloudinit/config/cc_scripts_user.pyc'>) failed#012Traceback (most recent call last):#012 File "/usr/lib/python2.7/dist-packages/cloudinit/stages.py", line 658, in _run_modules#012 cc.run(run_name, mod.handle, func_args, freq=freq)#012 File "/usr/lib/python2.7/dist-packages/cloudinit/cloud.py", line | 20:51 |
gimmic | 63, in run#012 return self._runners.run(name, functor, args, freq, clear_on_fail)#012 File "/usr/lib/python2.7/dist-packages/cloudinit/helpers.py", line 197, in run#012 results = functor(*args)#012 File "/usr/lib/python2.7/dist-packages/cloudinit/config/cc_scripts_user.py", line 38, in handle#012 util.runparts(runparts_path)#012 File "/usr/lib/python2.7/dist-packages/cloudinit/util.py", line 651, in runparts#012 % (len(failed), len( | 20:51 |
gimmic | attempted)))#012RuntimeError: Runparts: 1 failures in 1 attempted commands | 20:51 |
gimmic | small ui bug during deployment: https://i.imgur.com/RgPKPSq.png | 20:53 |
gimmic | system is ON, dropdown option is 'turn system on' which actually turns it off.. | 20:54 |
shewless | roaksoax/kiko: I tried to disable DHCP on all networks. I deleted the subnet entirely and then I re-added the subnet and re-enabled DHCP.. dhcpd still won't start because that file is missing | 20:56 |
shewless | what service is responsible for creating dhcpd.conf? I don't see any logs that indicate a problem there | 20:56 |
kiko | shewless, rackd does | 20:56 |
shewless | kiko: okay.. rackd doesn't want to tell me about any errors regarding dhcpd.conf I guess :( | 20:57 |
kiko | shewless, it seems to be confused about the IP address it needs to be listening on? | 20:57 |
shewless | kiko: interesting.. that might be the hint I needed | 20:58 |
jhobbs | /wg 6 | 20:59 |
shewless | kiko: so.. regiond.conf should be the "pxe" IP address right? | 21:00 |
shewless | kiko: It's working now.. I had it set to my external IP address instead somehow | 21:00 |
shewless | anyways.. dhcpd is running.. that's a good first step | 21:00 |
kiko | shewless, you have a single node running both rackd and regiond right? | 21:02 |
shewless | kiko: yup. and I think it's all working now (hopefully) just going to test deployment now | 21:03 |
kiko | shewless, cool | 21:04 |
kiko | shewless, this is an old bug, btw, that changing the MAAS server's IP address causes a mess | 21:05 |
kiko | gimmic, hmm | 21:05 |
gimmic | testing a theory, I changed the default filesystem to use LVM instead of flat | 21:06 |
gimmic | we'll see if deploys still err | 21:06 |
gimmic | i set one to allow ssh and got in to look | 21:06 |
gimmic | everything looks fine aside from that error I posted | 21:06 |
gimmic | seems if I deploy the same node twice, it deploys fine the second time | 21:06 |
gimmic | the raided node deployed fine | 21:12 |
mup | Bug #1523766 changed: Node stops commissioning and powers down before commissioning is complete <MAAS:Won't Fix> <https://launchpad.net/bugs/1523766> | 21:51 |
mup | Bug #1588547 opened: Generated bonding configuration is incorrect. <sts-needs-review> <MAAS:Triaged by mpontillo> <https://launchpad.net/bugs/1588547> | 22:21 |
mup | Bug #1588547 changed: Generated bonding configuration is incorrect. <sts-needs-review> <MAAS:Triaged by mpontillo> <https://launchpad.net/bugs/1588547> | 22:48 |
mup | Bug #1588547 opened: Generated bonding configuration is incorrect. <sts-needs-review> <curtin:New> <MAAS:Incomplete by mpontillo> <https://launchpad.net/bugs/1588547> | 23:46 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!