| alterjsive | is there any pxeboot friendly mode for port bonding? | 08:45 |
|---|---|---|
| alterjsive | wait let me rephrase that. Is there any port bonding mode that I can use during the installation of my servers without getting any faillures | 08:46 |
| alterjsive | active backup right? | 08:47 |
| alterjsive | gateway-connectivity: Failed to apply custom network configuration View log | 08:57 |
| alterjsive | Unable to communicate to the MAAS metadata service after applying custom network configuration | 08:57 |
| alterjsive | very wierd | 08:58 |
| alterjsive | port bonding doesn't work with ubuntu 20.04 and MAAS. I tried active backup. I get an error at the gateway-connectivity test: Failed to apply custom network configuration View log. Unable to communicate to the MAAS metadata service after applying custom network configuration. | 09:16 |
| alterjsive | Marking node failed - Node operation 'Deploying' timed out after 30 minutes. | 10:12 |
| alterjsive | This server has been installed without problems before. I don't understand why it's suddenly a problem | 10:32 |
| alterjsive | what is the best way to do a clean start? | 10:32 |
| alterjsive | maybe I need to delete all the partions? deleting the server in maas and pxe booting it again seems not enough | 10:33 |
| classicsnail | alterjsive: active-backup | 11:48 |
| alterjsive | classicsnail: I tried that, didn't work | 11:51 |
| alterjsive | I had to delete the bond to install the ubuntu servers | 11:51 |
| classicsnail | sounds like a switching configuration issue, the primary that comes up doesn't hvae a path for whatever reason | 11:52 |
| classicsnail | we use active-backup presently | 11:53 |
| alterjsive | weird, so if I would reset my switch settings to defaults. it should work? | 11:53 |
| classicsnail | active-backup simply detects a working path, based usually on ethernet link state - it doesn't need lacp or anything like that | 11:55 |
| alterjsive | yes, that's why I picked it | 11:55 |
| alterjsive | I have a 802.3ad router but I couldn't get that to work ether | 11:56 |
| alterjsive | same error | 11:56 |
| alterjsive | switch* | 11:56 |
| alterjsive | with LACP | 11:57 |
| classicsnail | pxe and lacp is odd, it depends entirely on the switch os if that's going to work | 12:12 |
| classicsnail | you can, but not every switch will support a tagged and untagged frame pathing | 12:13 |
| classicsnail | forwarding even | 12:13 |
| classicsnail | especially if they're the same vlan id | 12:13 |
| alterjsive | classicsnail: their all on the default vlan. | 12:34 |
| classicsnail | have you checked the console, or the syslog for the machine as delivered to maas? | 12:40 |
| classicsnail | there's usually an answer on either | 12:40 |
| classicsnail | all the extra services are working? such as dhcp and so on? | 12:40 |
| classicsnail | break it down to the simplest test you can do, and build up from there? install it using just a single link, no vlans or other trickery? | 12:40 |
| alterjsive | I'm trying to install the system without bond's first | 12:41 |
| alterjsive | for some reason it doesn't work anymore | 12:41 |
| alterjsive | Marking node failed - Node operation 'Deploying' timed out after 30 minutes. | 12:41 |
| alterjsive | I will check the logs next time it happens | 12:41 |
| alterjsive | if I can reach the machine | 12:41 |
| alterjsive | IRRC I couldn't ssh into the machine | 12:42 |
| alterjsive | got a bug here. When I have a machine that has failed in deployment and at the moment has a new state. And I set the dhcp server it's removed again when I start commisioning the machine. | 13:21 |
| alterjsive | I tried validate the configuration this time. | 13:22 |
| alterjsive | that seems to work | 13:22 |
| alterjsive | I will report it if I come across it again | 13:23 |
| alterjsive | classicsnail: isn't 802.3ad an IEEE standard? | 13:59 |
| antonym | hello, i'm running 2.9 snap of maas, is there a good quick way to debug the maas grub bootloader or a way to recompile it to edit options? maas commission is hanging at Booting under MAAS direction on certain broadcom cards on some newer platforms and it's hard to debug... i can get around it by forcing it to use iPXE with snponly.efi, but then find that the code does not create the proper /boot/efi | 16:53 |
| antonym | partitions because i think the iPXE code path assumes legacy boot only | 16:53 |
| ltrager | antonym: currently there isn't an easy way. You have to set 'debug="all"' in the grub template but those are stored in the snap and thus set to be read only. | 20:33 |
| ltrager | antonym: The next version of MAAS will enable GRUB debug mode if MAAS debug mode is enabled. | 20:33 |
| antonym | i think i found a way, i mounted an edited file outside of the snap to modify the grub template, was able to get into grub to debug now at least | 20:34 |
| antonym | was having issues with BCM7420s booting on a new Dell R7525, Intels seem to work ok but BCMs just seem to hang and stop | 20:42 |
| antonym | it looks like it can load up grub ok and the configs remotely, but with debug=all enabled, it fails to retrieve the kernels, fails on kern/disk.c:281: opening 'http,10.127.88.10:5248' failed | 21:51 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!