cpaelzer | bryceh: so your local build which isn't a full container more a chroot did change your systems time, that sounds wrong :-) Was that reproducible or didn't you try to not mess up the system more? | 06:15 |
---|---|---|
cpaelzer | as sdeziel said, for real containers you do not have CAP_SYS_TIME which some handle better than others, but if your env was really just a (s)chroot then your case isn't the same | 06:16 |
bryceh | cpaelzer, very reproducible. I created a second fresh lxc container and reproduced it there as well. | 06:53 |
cpaelzer | oh so it really was lxd container then | 06:57 |
bryceh | correct | 06:58 |
cpaelzer | bryceh: any chance we could replace ntpd in that env with chrony (the meant to work and supported ntp server) or at least (more a drop in replacement) ntpsec ? | 06:58 |
cpaelzer | the world is about to stop caring for classic ntpd in favor of one or the other | 06:58 |
bryceh | do you mean temporarily for testing, or permanently as in the autopkgtest config? | 06:59 |
bryceh | cpaelzer, I think my next step will be to redo the merge from scratch and see if it occurs again, or if I can spot where an error crept in. If that still repro's I'll post the branch for someone else to doublecheck. | 07:05 |
cpaelzer | sounds fine if it wasn't a 5 day effort that you are now re-doing :-) | 07:06 |
cpaelzer | and I meant temporarily for testing, depending on the outcome we can then make long term decisions | 07:06 |
=== slyon_ is now known as slyon | ||
=== jchittum_ is now known as jchittum | ||
=== jgee0 is now known as jgee | ||
=== lotuspsychje_ is now known as lotuspsychje | ||
=== haggertk_ is now known as haggertk | ||
=== kenyon_ is now known as kenyon | ||
ahasenack | good morning | 11:52 |
* kanashiro waves | 11:54 | |
=== jgee6 is now known as jgee | ||
ahasenack | quick non-scientific poll: do you use ansible on ubuntu server? Or is it too old? | 14:45 |
rbasak | I do. I use the version of ansible currently packaged in Focal. | 14:49 |
cpaelzer | ahasenack: I'm sure you've seen it, but there is feedback on the FRR MIR | 15:21 |
ahasenack | cpaelzer: yes, I saw it in the morning | 16:03 |
ahasenack | paride: https://pypi.org/project/ansible/ says latest ansible is 5.2.0, but the github project has "stable-2.X" branches | 16:28 |
ahasenack | and there is a split somewhere, ansible-core | 16:28 |
ahasenack | can you give a quick tl;dr? | 16:28 |
ahasenack | 2.12 seems to be the latest stable, is that? What is 5.2.0 then? | 16:28 |
ahasenack | https://groups.google.com/g/ansible-announce/c/Aw5eGIHgpkg <-- announcement for 2.12.1, ansible-core 2.11.7, ansible-base 2.10.16, just a few months old | 16:29 |
paride | ahasenack, it seems that they ramped up the versions numbers rather quickly https://groups.google.com/g/ansible-announce/search?q=release | 16:39 |
ahasenack | The homepage ansible.com is impossible to use :/ | 16:39 |
paride | ahasenack, the timeline in https://en.wikipedia.org/wiki/Ansible_(software) is useful but lacks sources | 16:45 |
paride | I mean, lacks references to official sources | 16:45 |
ahasenack | and meaning | 16:46 |
ahasenack | I see yellow for "extended support" in the legend, but no yellow in the graph | 16:46 |
ahasenack | maybe worth asking in their channel | 16:47 |
ahasenack | their #topic: " * latest releases: ansible 5.2.0, 2.9.27; ansible-core 2.12.2, 2.11.8; ansible-base 2.10.17; ansible 2.8 is EOL 💀; ansible 2.9 EOL announcement https://groups.google.com/g/ansible-announce/c/kegIH5_okmg/" | 16:47 |
paride | maybe https://docs.ansible.com/ansible/devel/reference_appendices/release_and_maintenance.html explains, reading... | 16:48 |
paride | ahasenack, actually the latest NEWS entry helps: https://salsa.debian.org/debian/ansible/-/blob/master/debian/NEWS | 16:56 |
ahasenack | is there another source in debian too? | 16:58 |
paride | yes, ansible-core | 16:59 |
paride | but ansible (2.10.7+merged+base+2.10.8+dfsg-1) they merged ansible-base (= ansible-core) into ansible, for some reason. | 17:02 |
ahasenack | aha, that's what "merged" means in that version | 17:06 |
ahasenack | sounds like debian is still experimenting with this | 17:07 |
ahasenack | hmml, that is not the experimental one even | 17:07 |
ahasenack | it's a huge package, over 200Mb installed | 17:21 |
ahasenack | looks like it has tests installed as well | 17:21 |
ahasenack | paride: I'm finding that many fixes that were in ansible-core (deb) were not pulled into the new merged ansible package | 17:52 |
ahasenack | maybe they were just in experimental at that time, could be | 17:52 |
znf | ahasenack, there's a reason ansible folks recommend using ansible from pip :) | 18:54 |
ahasenack | pip has 5.2, which from my understanding is a bundle | 18:55 |
ahasenack | the core/base bits, and some "curated" playbooks I think | 18:55 |
rbasak | znf: that's a really common recommendation from most upstreams. Yet most of us use most of our software from distribution packages :) | 18:57 |
znf | some packages have a very fast release/development cycle (like ansible) | 18:59 |
znf | and there's a certain point you kinda want the new features :) | 19:00 |
rbasak | Every two years, when there's a new LTS out. Then it's on a nice, predictable cadence :) | 19:02 |
rbasak | Then it's nicely in phase with the support cycles of the machines I'm actually managing, too. The last thing I want is to find some issue with the playbook of an existing deployment that worked previously because something did a major version bump without coordination, or for me to be unable to reproduce a setup that worked previously. | 19:04 |
sergiodj | kanashiro: FWIW https://code.launchpad.net/~sergiodj/britney/+git/hints-ubuntu/+merge/414905 | 19:18 |
kanashiro | sergiodj, thanks! | 19:19 |
sergiodj | after that's processed, there'll be only jekyll | 19:20 |
kanashiro | I have jekyll almost fixed locally | 19:38 |
=== genii is now known as genii-core |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!