| holmanb | meena: how does this look? https://github.com/canonical/cloud-init/pull/2143 | 16:57 |
|---|---|---|
| -ubottu:#cloud-init- Pull 2143 in canonical/cloud-init "Fix linux-specific call on BSD." [Open] | 16:57 | |
| holmanb | meena: also thanks for the instructions, worked like a charm | 16:58 |
| minimal | has any thought ever been given to cloud-init running using pyc files only? | 17:05 |
| meena | holmanb: that exception has become background radiation; It was on my todo list, but i forgot it's something fixable, and not just part of nature | 17:08 |
| meena | minimal: does that even work? | 17:09 |
| meena | minimal: the problem might be that pyc is very very very platform specific | 17:09 |
| holmanb | meena: Yeah it didn't seem high prior, but it looked like a quick fix so here we are :) | 17:10 |
| minimal | meena: the pyc files are created for the platform when cloud-init is "built" | 17:15 |
| meena | holmanb: I'm looking if i can fix mountinfo | 17:16 |
| meena | the amount of stuff i run into because i can't upgrade my 14.0-CURRENT VMs… | 17:32 |
| meena | i need to get pkgbase.live up again | 17:33 |
| meena | this is mount | grep tmp: on my Linux desktop: https://gist.github.com/igalic/fdaff6da508f030ecfde48ee2d9543d0 | 17:37 |
| meena | and, wowie wow, i don't think I've kept up with linux for quite some time | 17:37 |
| meena | holmanb: I'll try to fix the root cause of that noexec parsing, cuz, we do have have that flag | 18:15 |
| meena | oooooohh | 18:18 |
| meena | mount has a --libxo! | 18:18 |
| holmanb | meena: sounds good, however you prefer | 18:38 |
| meena | looks like I should've shutdown my vm, before taking a snapshot… all the modifications I made to the image are lost | 18:53 |
| meena | i wish i had ZFS on this computer lol | 18:53 |
| meena | so, how does the LXD DataSource work? | 18:55 |
| meena | do i need the LXD Agent for that? | 18:55 |
| meena | given the results I'm getting from Cloud-Init, I'd say: yes. | 19:07 |
| meena | i love yak shaving | 19:08 |
| meena | it's my favourite thing to do | 19:08 |
| blackboxsw | meena: for what it's worth for older versions of Ubuntu, we had to setup LXD profiles that automatically setup lxd-agent in some images https://github.com/canonical/pycloudlib/blob/main/pycloudlib/lxd/defaults.py#L9 | 20:46 |
| blackboxsw | and yes DataSourceLXD requires the lxd-agent I think which sets up /dev/lxd/sock file | 20:46 |
| === esv_ is now known as esv | ||
| meena | blackboxsw: https://github.com/lxc/lxd/issues/11603 | 20:53 |
| -ubottu:#cloud-init- Issue 11603 in lxc/lxd "compiling lxd-agent for FreeBSD" [Open] | 20:53 | |
| meena | stgraber says it's easy | 20:53 |
| meena | holmanb: all that temp_utils code is really confusing. | 22:08 |
| meena | on what grounds does this happen: | 22:09 |
| meena | In [56]: temp_utils.get_tmp_ancestor(needs_exe=True) | 22:09 |
| meena | Out[56]: '/var/tmp/cloud-init' | 22:09 |
| meena | In [57]: temp_utils.get_tmp_ancestor(needs_exe=False) | 22:09 |
| meena | Out[57]: '/tmp' | 22:09 |
| meena | that function is just reading random environment variables to make those decisions | 22:09 |
| meena | neither on my Linux laptop, nor on my FreeBSD VMs is there any functional difference between /var/tmp and /tmp | 22:10 |
| meena | aah, but there is, on Linux, if I run it as root | 22:11 |
| meena | tmpfs on /run type tmpfs (rw,nosuid,nodev,noexec,relatime,size=6572688k,mode=755,inode64) | 22:11 |
| meena | we don't have a /run dir on FreeBSD. We only have /var/run, and it's not mounted as noexec. | 22:12 |
| meena | almost done… | 23:31 |
| meena | https://github.com/canonical/cloud-init/pull/2146 hrm… i based this on the wrong branch | 23:38 |
| -ubottu:#cloud-init- Pull 2146 in canonical/cloud-init "FreeBSD fix parsing of mount and mount options" [Open] | 23:38 | |
| meena | either way, that commit is safe. getting that regex was hard work lol | 23:38 |
| meena | tomorrow I'll extract the path parsing from parse_mount_info() | 23:38 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!