| mup | Bug #1771885 changed: bionic: static maas missing search domain in systemd-resolve configuration <bionic> <network> <cloud-init:Won't Fix> <juju:Fix Committed by ecjones> <juju 2.3:Fix Released by ecjones> <MAAS:Invalid> <https://launchpad.net/bugs/1771885> | 00:29 |
|---|---|---|
| enrico_ | hello, please hope you can help, I have deployed a maas cluster, I am able to manage power control of a node that I added manually on maas | 15:46 |
| enrico_ | the node is well commissioned | 15:46 |
| enrico_ | I want do deploy ubuntu on top | 15:46 |
| enrico_ | it seems stucked at: curtin: Installation started | 15:47 |
| enrico_ | cloud-init third party drivers not installed or necessary | 15:47 |
| enrico_ | and nothing more | 15:47 |
| enrico_ | maybe it does not arrive to use maas images ? | 15:47 |
| enrico_ | can you help to debug and figure out the problem please? | 15:48 |
| mup | Bug #1775424 opened: CentOS 6 deployment fails with AttributeError: 'NoneType' object has no attribute 'groups' <MAAS:New> <https://launchpad.net/bugs/1775424> | 16:07 |
| roaksoax | enrico_: can you share the rsyslog for the machine ? /var/log/maas/rsyslog/<machine-name>/<date>/messages | 16:07 |
| xygnal_ | roaxsoax; | 16:30 |
| xygnal_ | only tried in dev so far | 16:31 |
| enrico_ | roaksoax https://pastebin.com/kExGTtS2 | 16:41 |
| enrico_ | these are tail -100 messages | 16:41 |
| roaksoax | xygnal_: k, let me know how it goes | 16:41 |
| mup | Bug #1775430 opened: Make gw optional for static routes <canonical-bootstack> <networking> <MAAS:New> <https://launchpad.net/bugs/1775430> | 16:43 |
| roaksoax | enrico_: uhmm there's no real error surfaced at all | 16:44 |
| roaksoax | enrico_: what's the status of the machine while you see this? 'Deploying'? | 16:45 |
| enrico_ | yes Deploying... and isntallation output still System is booting... | 16:45 |
| roaksoax | enrico_: yeah that's because it never failed, and what happens if you leave it ? | 16:48 |
| enrico_ | if I dont touch it is seems to be in a stuck state forever ( I waited for more than 30min ) | 16:48 |
| enrico_ | If I release the node.. the result is failed deployment | 16:49 |
| roaksoax | enrico_: looks to me like curtin gets stuck | 16:49 |
| enrico_ | trying to redeploy again to generate new log | 16:50 |
| enrico_ | yes it could be but dont know why | 16:50 |
| enrico_ | you have any idea? | 16:50 |
| roaksoax | enrico_: ssh into the machine while dpeloying | 16:51 |
| roaksoax | touch /tmp/block-reboot | 16:51 |
| roaksoax | enrico_: and tail /var/log/cloud-init.log | 16:52 |
| roaksoax | and see if there are any errors | 16:52 |
| enrico_ | ok | 16:52 |
| enrico_ | 2018-06-06 16:49:22,334 - handlers.py[DEBUG]: finish: modules-final/config-scripts-per-instance: SUCCESS: config-scripts-per-instance ran successfully 2018-06-06 16:49:22,335 - url_helper.py[DEBUG]: [0/1] open 'http://55.0.51.31:5240/MAAS/metadata/status/npnec6' with {'method': 'POST', 'allow_redirects': True, 'url': 'http://55.0.51.31:5240/MAAS/metadata/status/npnec6', 'headers': {'Authorization': 'OAuth oauth_nonce="1746154607730300 | 16:53 |
| enrico_ | https://pastebin.com/NhKAtHNz | 16:53 |
| roaksoax | seems likes running just fine | 16:54 |
| roaksoax | enrico_: dpkg -l | grep cloud-init | 16:54 |
| roaksoax | to get the version | 16:54 |
| enrico_ | ii cloud-init 18.2-4-g05926e48-0ubuntu1~16.04.2 all Init scripts for cloud instances | 16:54 |
| enrico_ | ii cloud-initramfs-copymods 0.27ubuntu1.5 all copy initramfs modules into root filesystem for later use | 16:54 |
| enrico_ | ii cloud-initramfs-dyn-netconf 0.27ubuntu1.5 all write a network interface file in /run for BOOTIF | 16:54 |
| roaksoax | enrico_: can you try to refresh your maas images and see if it downloads updated images | 16:55 |
| roaksoax | seems a new version of cloud-init is already released | 16:55 |
| roaksoax | on may 16th, which should be on the latest images | 16:55 |
| enrico_ | I have installed maas after may 16th | 16:56 |
| enrico_ | I use play maas over ubuntu 16 | 16:56 |
| enrico_ | cloud-init should be installed on maas rack/region controller right? | 16:59 |
| roaksoax | enrico_: no, it comes from the maas images themselve | 17:00 |
| roaksoax | enrico_: so go to the images tab and click 'Update selection' on the ubuntu section | 17:01 |
| enrico_ | when you say: maas images... you refer to ubuntu ISO that includes maas ? | 17:02 |
| enrico_ | ok | 17:02 |
| roaksoax | enrico_: In the 'Images' tab | 17:02 |
| roaksoax | enrico_: the images maas downloads | 17:02 |
| enrico_ | yes updated | 17:02 |
| roaksoax | enrico_: ok, so try agian and check the cloud-init version to see if that has changed | 17:05 |
| enrico_ | yes done.. it is the same | 17:06 |
| roaksoax | uhmm strange, so i winder if its a cloud-init issue | 17:07 |
| roaksoax | i'll dig and ask around | 17:07 |
| enrico_ | Thanks you for your time.. I appreciate :) | 17:07 |
| enrico_ | If you discover something please contact me on enricodhd@gmail.com | 17:08 |
| enrico_ | anyway I'll ping you in a while here | 17:08 |
| xygnal_ | roaksoax: whats the max latency/distance between rack controllers and region controllers, and how might it impact maas if we had a region controller in another country than a rack controller. | 17:13 |
| roaksoax | xygnal_: well in most cases should be just fine, depending how many machines you may want to deploy at the same time | 17:14 |
| roaksoax | xygnal_: i guess the major bottleneck is tftp because tftp sucks, which could mean if you deploying several machines at the same time they would take a long time to donwload the initrd/kernel | 17:15 |
| roaksoax | but other than that everything should be ok, since its http | 17:15 |
| xygnal_ | we have two rack we want to deploy to India but them to test and play in, but it would not see a great deal of builds itself | 17:15 |
| xygnal_ | how would it handle the delays in check in? | 17:15 |
| xygnal_ | this would be literally across the globe | 17:15 |
| roaksoax | xygnal_: 2 racks, say 10 machines? 20 machines ? | 17:17 |
| xygnal_ | about 16 maciens | 17:21 |
| xygnal_ | machines* | 17:21 |
| roaksoax | xygnal_: yeah so the only "bottleneck" i could see if if you deploy the 16 machines at once, whether they would be able to download the initrd/kernel over tftp fast enough | 17:23 |
| roaksoax | xygnal_: but technically, shouldn't really be a problem | 17:23 |
| === Guest55849 is now known as jefferai | ||
| === jefferai is now known as Guest18839 | ||
| === Guest18839 is now known as ferai | ||
| xygnal_ | tftp would be coming from the region? | 17:28 |
| xygnal_ | not the rack? | 17:28 |
| roaksoax | xygnal_: rack | 17:30 |
| xygnal_ | rack would be in same place at those 16 blades, its region that would be far away. | 17:33 |
| xygnal_ | why would tftp be slow? | 17:33 |
| roaksoax | xygnal_: you should be fine then | 17:34 |
| === ferai is now known as jefferai | ||
| === jefferai is now known as Guest50383 | ||
| bdx | roaksoax: could you advise on how I might get out of this debacle with the stuck controller? | 18:54 |
| === Guest50383 is now known as jefferai | ||
| xygnal_ | roaksoax: I can't seem to find evidence of a CentOS 7 package for Cloud-init newer then 0.7.x :/ ever done a custom build/install of it? | 19:20 |
| xygnal_ | roaksoax: 0.7.9-24 is the latest in theirs repos | 19:21 |
| xygnal_ | cloud-init | 19:21 |
| mup | Bug #1775461 opened: manual node inaccessible after commissioning <MAAS:New> <https://launchpad.net/bugs/1775461> | 20:07 |
| mup | Bug #1770201 changed: DNS resolution issues during enlistment with Bionic ephemeral environment. <cdo-qa> <cdo-release-blocker> <foundations-engine> <MAAS:Invalid> <https://launchpad.net/bugs/1770201> | 22:13 |
| mup | Bug #1774666 changed: Bond interfaces stuck at 1500 MTU on Bionic <cdo-qa> <foundations-engine> <mtu> <netplan> <cloud-init:New> <MAAS:Invalid> <cloud-init (Ubuntu):Confirmed> <netplan.io (Ubuntu):Confirmed> <https://launchpad.net/bugs/1774666> | 23:47 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!