mup | Bug #1687420 opened: Maas api /machines/{system_id}/?op=restore_networking_configuration does not clear network config <MAAS:New> <https://launchpad.net/bugs/1687420> | 14:19 |
---|---|---|
vasey | mpontillo: i have a node (manual power control) i want to use for my juju controller node, but the RAM is showing up as 0.0 GB...any idea why that might be? | 14:55 |
vasey | mpontillo: update, i'm now using the vmware power control successfully (followed your certificate fix steps), and i'm still seeing 0.0GiB RAM, even though it's configured with 3.5GB | 15:28 |
vasey | mpontillo: and the juju bootstrapper needs to see in maas that the system has 3.5GB of ram for it to deploy, but it's just not there atm | 15:29 |
wasosa | Hi all, I'm trying to get going with maas, but I'm stuck on getting storage information during commissioning. The 00-maas-07-block-devices.out I get back looks ok to me, but I cannot deploy because storage info is 'missing'. Any help is appreciated. I'm using maas 2.1.5+bzr5596-0ubuntu1~16.04.1. | 15:36 |
wasosa | I'm trying to deploy to a machine with an nvme drive; I found this https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1642903, but I'm not sure that's the issue (symlinks look ok on the node). | 15:37 |
xygnal | roaksoax: mpontillo: estimates on rc4? waiting on thoe missing hooks for dhcp relay. | 15:42 |
pmatulis | vasey, did you commission it? | 15:43 |
vasey | pmatulis: yes, just deployed ubuntu 17 to it as well, still isn't showing any RAM in the maas ui | 15:43 |
roaksoax | xygnal: in progress, mpontillo will work on those | 15:56 |
vasey | pmatulis: trying to add the machine manually now | 16:03 |
vasey | pmatulis: still no dice ;( is this a vmware issue or possibly a lshw output formatting issue? is there a way to manually tell MAAS how uch RAM a machine has? | 16:22 |
wasosa | I can see where the errors are added(storage_layout_issues() in /usr/lib/python3/dist-packages/maasserver/models/node.py), but I haven't yet figured out where the problem occurs. | 16:33 |
vasey | mpontillo: pmatulis: https://pastebin.com/ff3DE6aP my 00-maas-01-lshw output contains the amount of ram my VM has; why isn't it being reported in the maas machine description? | 16:35 |
=== stormmore is now known as Budgie^Smore | ||
mpontillo | vasey: can you post any relevant details on this bug report? https://bugs.launchpad.net/maas/+bug/1682150 | 16:48 |
mpontillo | vasey: if you can pastebin full output from your lshw commissioning script it may help us to write a test case | 16:50 |
vasey | mpontillo: thanks, here's that output https://pastebin.com/upD20kWv | 17:19 |
bigtexun | I changed the IP address of my MAAS server as I was learning how to configure it, and I ran sudo dpkg-reconfigure maas-region-controller, but when I deploy a machine MAAS inserts the old IP address in resolv.conf, where do I fix that? | 17:46 |
bigtexun | It also inserts the new IP... But DNS is timing out on the deployed machine, so things like sudo take forever. Deployment was successful, despite MAAS failing the deployment. I assume the DNS timeout is triggering the deploy fail indication in MAAS... | 17:48 |
wasosa | I was able to confirm that `self.blockdevice_set.all()` returns an empy list (via print statement) inside storage_layout_issues(). I still can't tell where that comes from, but I'll keep digging. Any pointers are welcome ;-) | 17:58 |
mpontillo | vasey: that's interesting, but I was actually interested in the full output from the 00-maas-01-lshw commissioning script. looks like you pasted the lldp output | 18:19 |
mpontillo | bigtexun: check /etc/maas/regiond.conf | 18:20 |
mpontillo | bigtexun: and /etc/maas/rackd.conf | 18:20 |
mpontillo | bigtexun: it sounds like you've reconfigured the region but may also need to reconfigure the rack; try "sudo dpkg-reconfigure -plow maas-rack-controller" | 18:21 |
vasey | mpontillo: sorry about that! here it is: https://pastebin.com/01kV39Ee | 18:25 |
wasosa | One more crumb: I only see one row in the maasserver_blockdevice table, and it doesn't look like my drive:maasdb=# select * from maasserver_blockdevice; id | created | updated | name | id_path | size | block_size | tags | node_id ----+-------------------------------+-------------------------------+------+----------------------------------------+- | 18:38 |
wasosa | Bah, sorry. | 18:39 |
wasosa | id | created | updated | name | id_path | size | block_size | tags | node_id ----+-------------------------------+-------------------------------+------+----------------------------------------+--------------+------------+------------+--------- 1 | 2017-04-19 13:37:46.604938-07 | 2017-05-01 10:54:26.505265-07 | sda | /dev/disk/by-id/wwn-0x5002 | 18:39 |
pmatulis | wasosa, use a pastebin next time | 19:09 |
wasosa | Will do, thanks: https://pastebin.com/ZB4Ef3ML | 19:13 |
wasosa | So it seems the blockdevice info is coming back from the node to the server, but is not making its way into the database. | 19:14 |
xygnal | question: when does the user_data metadata get passed through to cloud-init on the guest? | 19:20 |
xygnal | is after after deployment? during deployment? | 19:20 |
xygnal | one follow up question to that one, is user_data pased through regardless of image source?we use custom, but our image has cloud-init | 19:43 |
xygnal | would like to apply the user_data field | 19:44 |
bigtexun | mpontillo: Thanks! I'll try that. Shortly before you posted I decided to try rebooting the MAAS server, and that caused PXE to start failing, so perhaps this will fix me up... | 19:52 |
roaksoax | bigtexun: dpkg-resource the rack controller too | 19:56 |
Mac_ | Hi there, | 20:00 |
Mac_ | I'm running CiaB to deploy Thunder ARM64 compute node by MAAS, and I have had no problem with that until now. This week when I tried to do it again, the Thunder alway give me the following message: | 20:00 |
Mac_ | CORD cord-in-a-box script | 20:01 |
Mac_ | EFI stub: Booting Linux Kernel... ConvertPages: Incompatible memory types EFI stub: ERROR: Failed to alloc kernel memory EFI stub: ERROR: Failed to relocate kernel Unloading driver at 0x11F8400A000 Failed to boot both default and fallback entries. Press any key to continue... Booting under MAAS direction... error: couldn't send network packet. unaligned pointer 0x11fea8f89d0 borted. Press any key to exit.error: you need to load the ke | 20:01 |
Mac_ | MAAS Version 1.9.5+bzr4599-0ubuntu1 (14.04.1) This seems to me that something wrong with the EFI image MAAS has. Any suggestions for further debugging? Thanks. | 20:01 |
bigtexun | roaksoax: Thanks, I only have the one server... This is a first-time installation. Everything except DNS was working on the deployed node this morning... I was focused on just fixing the IP problem in DNS, and I rebooted after I posted my question today... And after booting pxe stopped working. | 20:02 |
bigtexun | Since I have no data, just my MAAS sandbox, and a single physical node, I'm going to do a clean install with no IP address changes and see if everything works. | 20:06 |
bigtexun | perhaps I fat-fingered something in the week I spend getting this thing working. | 20:06 |
mup | Bug #1687463 opened: Pods listing page header still shows "Add pod" when pod from list is selected. <MAAS:In Progress by newell-jensen> <MAAS RSD :In Progress by newell-jensen> <https://launchpad.net/bugs/1687463> | 20:14 |
roaksoax | Budgie^Smore: restarting regiond and rackd | 20:31 |
roaksoax | err | 20:31 |
roaksoax | sry | 20:31 |
roaksoax | /w/win 5 | 20:41 |
Budgie^Smore | lol no worries roaksoax :) | 21:05 |
mup | Bug #1687487 opened: Virsh pod creation failed with "Node with hostname already exists." <MAAS:Confirmed> <https://launchpad.net/bugs/1687487> | 22:35 |
mup | Bug #1687500 opened: Unable to configure DHCP due to IPAddress not assoicated with subnet <MAAS:In Progress by ltrager> <https://launchpad.net/bugs/1687500> | 23:35 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!