20:01:16 <briantist> #startmeeting Ansible Windows Working Group
20:01:16 <zodbot> Meeting started Tue Apr 27 20:01:16 2021 UTC.
20:01:16 <zodbot> This meeting is logged and archived in a public location.
20:01:16 <zodbot> The chair is briantist. Information about MeetBot at http://wiki.debian.org/MeetBot.
20:01:16 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
20:01:16 <zodbot> The meeting name has been set to 'ansible_windows_working_group'
20:01:23 <briantist> lol I didn't think that would work
20:03:57 <briantist> I'm not sure if anyone is actually here, or if this meeting was skipped or whatever.. was not intending to actually run it, but I guess I'll stick around for a few just in case and then end it unless an actual chair shows up
20:06:00 <briantist> nothing on the agenda https://github.com/ansible/community/issues/581
20:06:28 <briantist> #topic Open Floor
20:09:15 <jborean93> sorry time slipped past me
20:09:28 <briantist> #chair jborean93
20:09:28 <zodbot> Current chairs: briantist jborean93
20:09:37 <jborean93> thanks for kicking it off though briantist
20:09:39 <jborean93> how are you?
20:09:40 <briantist> that's ok, usually happens to me too
20:09:58 <briantist> hehe I honestly didn't think I could, was kind of doing it jokingly 😅
20:10:24 <briantist> I'm alright, been hectic hence why I missed a few
20:10:30 <briantist> yourself?
20:10:43 <jborean93> not too bad, have too many things I need to comlpete and not enough time
20:10:55 <briantist> I know that tune...
20:11:05 <jborean93> I've resigned myself to the fact I need to touch win_updates which is not my favourite thing in the world
20:11:13 <briantist> 😭😭😭
20:11:17 <briantist> what's going on with it?
20:11:36 <jborean93> a few issues have popped up that I've been somewhat ignoring
20:12:10 <jborean93> I'm also in the middle of trying to make win_reboot a bit more stable with psrp and ssh
20:12:33 <jborean93> Currently going through the rounds of testing it out as the issues I'm trying to solve are quite sporatic in nature
20:12:52 <briantist> we are using `win_updates` but it's being phased out for a non-ansible update system, and unfortunately we don't monitor the ansible one well, so I'm not even sure if it's misbehaving for us
20:13:21 <briantist> oh does `win_reboot` have issues with `psrp`? what happens?
20:14:12 <jborean93> I've encountered a few problems especially when it comes to a reboot with a domain join. But I'm unsure if it's psrp specific or if it also happens with WinRM
20:14:47 <briantist> does it look vaguely like a DNS issue by any chance?
20:14:51 <jborean93> It's just unfortunate I missed a few exceptions in exec_command so some request timeout (Connect/Read Timeout) are raised and not converted to an AnsibleConnectionError that win_reboot sometimes catches
20:15:44 <briantist> #chair nitzmahone
20:15:45 <zodbot> Current chairs: briantist jborean93 nitzmahone
20:15:49 <jborean93> hey
20:15:55 <briantist> howdy
20:16:22 <nitzmahone> sorry, IRCCloud hassles
20:16:40 <nitzmahone> Nothing exciting from me
20:18:19 <briantist> jborean93: the reason I asked about DNS, is I saw a strange issue last year during a domain join process on EC2 instances, where a `Restart-Computer` call would fail with a very strange looking DNS resolution error, somewhere deep in the stack, but I couldn't actually show DNS resolution issues with any other tool, and `shutdown -r -t 0` worked.
20:18:22 <jborean93> here's the draft PR I thought I opened https://github.com/ansible-collections/ansible.windows/pull/214
20:18:51 <jborean93> was it a ps stack or python stack?
20:19:00 <briantist> but we could only surface that problem when assigning DNS servers via DHCP. Super strange (nothing to do with ansible btw)
20:19:09 <briantist> ps
20:19:26 <jborean93> we also do `-t 2` (or maybe 5) to give it time to return the output before it closes the connection
20:19:46 <jborean93> yea the errors are more Python requests based errors that I sometimes saw slip through
20:20:37 <briantist> it was the userdata script run by AWS on startup, we'd been using the same script for a long time. But we were using `Restart-Computer` cmdlet, and it always worked. Just started failing when we switched to DHCP DNS assignment instead of static. Never did find an explanation as a reorg happened and the DHCP/DNS project was suspended
20:20:48 <briantist> got it ok
20:21:23 <jborean93> weird maybe it still tries to resolve the local hostname even when `-ComputerName` isn't specified
20:22:11 <jborean93> in any case I need to try these changes out a bit more before merging. Once that is done it's onto win_updates which will hopfully utilise the new plugin_util instead of calling win_reboot directly
20:22:12 <nitzmahone> Just out of curiosity, were you renaming the host at the same time with the domain-join? I've seen more problems from that case than when the hostname stays the same
20:22:13 <briantist> yeah that's the weird thing though, running other commands (`nslookup`, .net object DNS resolution, cmdlets, etc.`) run before that call all resolve the computer name, domain name, `localhost` successfully.
20:22:52 <briantist> yeah definitely it was being renamed, but that part has always worked flawlessly (still does)
20:23:28 <briantist> anyway that's all ancient history for me at this point
20:23:36 <briantist> jborean93: will check out that PR
20:24:09 <nitzmahone> +1 on that PR as well (conceptually, anyway)- the subclassing made more sense when everything was in core, but scary when they're split
20:24:38 <jborean93> yea nothing has broken but I don't want to risk it
20:25:52 <briantist> "Removed ``shutdown_timeout`` and ``shutdown_timeout_sec`` which has not done anything since Ansible 2.5"
20:26:20 <briantist> they are still listed in valid args, is that to not breka backwards compat?
20:26:43 <jborean93> they aren't documented
20:26:49 <jborean93> and haven't done anything since 2.5
20:26:52 <briantist> ahhh ok
20:27:07 <jborean93> it also had a warning message whenever they were used, I think it's about time to remove them
20:27:37 <briantist> ok yeah I was just thinking if they were official they should be breaking changes/major version, but if they were never documented seems fine to remove by now
20:27:40 <jborean93> while we never said in the message when it was going to be removed I would hope that nobody has it set anymore
20:28:28 <briantist> 👍
20:28:29 <jborean93> if they do then really I prefer them to just remove the options as anybody who uses this collection will be on at least 2.9
20:31:27 <briantist> well, nothing else from me
20:34:04 <jborean93> cool, I've got nothing else to add as well
20:35:37 <briantist> sounds good, thanks, sorry for stealing a chair haha ;)
20:35:42 <briantist> #endmeeting