nils_ | am I correct in assuming that users are created before disk-setup is run? If I needed to put the home directories for users on a different device then I should have the device mounted and formatted in bootcmd? | 07:23 |
---|---|---|
felipe_1982_ | I am running bookhook scripts in AWS, and the "echo" is not getting logged in /var/log/cloud-init-output.log as the documentation would suggest. Running Amazon Linux 2. | 08:03 |
Odd_Bloke | rharper: Hmm, I was going to suggest cc_growpart/cc_resizefs for the partition expansion question from yesterday, but it looks like cc_resizefs _only_ resizes the root FS. | 11:46 |
lfain | What is the recommended way to access vendor data in a pre-installed cloud-init script? Right now, I get it by calling http://169.254.169.254/openstack/2018-08-27/meta_data.json but I'm looking for a generic method that will work with any datasource type. | 13:03 |
Odd_Bloke | lfain: o/ What do you mean by a "pre-installed cloud-init script"? You may want to look at `cloud-init query`. (Run `cloud-init query -a` to see all the instance-data.) | 14:08 |
rharper | Odd_Bloke: yeah, it's always been root volume only; | 14:17 |
rharper | nils_: look at /etc/cloud/cloud.cfg , 'cloud_init_modules' shows the order in which they run; users-groups is next to last, and runs after disk_setup and mounts; which you'll need both to create the disks/partitions to hold the user dirs and then mount entries to mount before the user is created. | 14:19 |
nils_ | rharper, it may be the ancient version bundled with CentOS, in my case users-groups is last in cloud-init-modules, mounts is first in cloud-config-modules | 14:23 |
Odd_Bloke | rharper: Right, but growpart supports non-root volumes, so the discrepancy caught me off-guard. | 14:24 |
Odd_Bloke | I guess at _first_ boot, you generally won't have filesystems on anything but your root volume. | 14:24 |
rharper | nils_: that's ... odd | 14:24 |
Odd_Bloke | tribaal: https://bugs.launchpad.net/cloud-init/+bug/1839854 <-- as a CloudStack(-ish) operator, I'd love your thoughts on this bug | 14:25 |
ubot5 | Ubuntu bug 1839854 in cloud-init "CloudStack provider cannot determine correct metadata IP with multiple network interfaces" [Undecided,New] | 14:25 |
rharper | Odd_Bloke: it's possible that secondary disks could be partial, but unlikely; the resize was mostly about using cloud images overtop a later device | 14:25 |
lfain | I made a custom image where I put a script in the /var/lib/cloud/scripts/per-instance/ folder. This script uses parameters that I provide via the meta data. | 14:25 |
rharper | lfain: you may want your script to read /run/cloud-init/instance-data.json then; that has the instance metadata available | 14:26 |
lfain | rharper: I forgot to write that the image is based on CentOS7 with cloud-init v.18.2 installed. Therefore, I can not run "cloud-init query" and the /run/cloud-init/instance-data.json also is not available. | 14:39 |
rharper | hrm | 14:40 |
lfain | Is there a way to install latest cloud-init on CentOS7? I didn't find RPMs for it. | 14:40 |
rharper | https://copr.fedorainfracloud.org/coprs/g/cloud-init/cloud-init-dev/ we do have daily rpm builds; | 14:40 |
rharper | https://copr.fedorainfracloud.org/coprs/g/cloud-init/el-testing/packages/ | 14:41 |
rharper | has 19.2 | 14:41 |
lfain | Thank you, rharper. I'll try 19.2. If I run into a problem, then I will return back to my current method (that I don't like...) | 14:45 |
rharper | lfain: cool! | 14:48 |
=== lfain1 is now known as lfain | ||
AnhVoMSFT | Hi folks, I have a CI build failed for this MP https://code.launchpad.net/~vtqanh/cloud-init/+git/cloud-init/+merge/369785 but the link does not work for me so I can't see what went wrong | 17:04 |
rharper | AnhVoMSFT: let me see and I'll let you know | 17:51 |
AnhVoMSFT | thanks rharper | 18:01 |
rharper | AnhVoMSFT: looks like a harness issues, I'll rerun the job | 18:06 |
AnhVoMSFT | thanks | 18:08 |
AnhVoMSFT | rharper I got disconnected from IRC but it looked like CI build failed again | 18:34 |
rharper | AnhVoMSFT: yes, tracking issue down on our side | 18:34 |
rharper | I've poked at jenkins; so this next run should work; but you won't need to do anything; we'll make sure it runs to completion , and if there's an issue in the branch, I'll let you know | 18:35 |
Odd_Bloke | rharper: blackboxsw: What is the relation between /v/l/cloud/instance/user-data.txt and user-data.txt.i? | 18:46 |
AnhVoMSFT | thanks rharper | 18:52 |
rharper | Odd_Bloke: I'd have to look, I thought it was something to do with mime wrappers | 18:59 |
rharper | ah, userdata_raw => user-data.txt , "processed user-data" => user-data.txt.i | 19:00 |
rharper | Odd_Bloke: see cloudinit/stages.py:def _store_userdata() | 19:00 |
AnhVoMSFT | looks like it passed this time | 19:08 |
rharper | AnhVoMSFT: yeah, I'll review | 19:18 |
Odd_Bloke | rharper: Did you get to the bottom of that mkdir cloud_tests error? | 19:21 |
rharper | Odd_Bloke: yes; lxd snap on the host needed someone to sudo mkdir a path that we use | 19:23 |
rharper | I suspect the upgrading/updating may have removed it from the common dir | 19:23 |
Odd_Bloke | Aha, OK, thanks! | 19:23 |
Odd_Bloke | Glad I asked, else I was going to spend 3 hours on triage today. :p | 19:24 |
powersj | O.o | 19:25 |
blackboxsw | rharper: Odd_Bloke care if I publish cloud-init to eoan again (now that Azure v2 networking is landed | 20:16 |
blackboxsw | or shall we wait on OCI | 20:17 |
blackboxsw | Odd_Bloke: LGTM https://code.launchpad.net/~daniel-thewatkins/cloud-init/+git/cloud-init/+merge/371053 minor comments/questions for clarification but nothing blocking | 21:20 |
Odd_Bloke | blackboxsw: Good points, I'll fix those up first thing tomorrow. | 21:42 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!