[08:11] morning === sdeziel_ is now known as sdeziel [18:10] sdeziel: you around? [18:10] teward: I am [18:11] sdeziel: regarding bug #2081308 that sounds more like a problem with policy-rc.d or some other issue than the nginx package but you know more about policy-rc.d than I do, is this an nginx bug or a user-has-failed situation? [18:11] -ubottu:#ubuntu-server- Bug 2081308 in nginx (Ubuntu) "open() '/etc/nginx/nginx.conf' failed (2: No such file or directory)" [Undecided, Incomplete] https://launchpad.net/bugs/2081308 [18:11] because it looks like their issue is they're changing policy-rc.d and it's breaking rather than an actual nginx package issue [18:12] teward: that's a good question. I find the error report (missing `nginx.conf`) weird. How is that file's existence tied to the service starting or not is what I couldn' [18:12] t figure [18:13] I know that bryyce has a request to include /etc/nginx/nginx.conf in the apport hooks, but that error sounds more like whatever underlying code is executed during setup is busted. There is a note in Debian right now that `nginx` doesn't predepend on nginx-common which COULD suggest they somehow installed `nginx` without `nginx-common` [18:14] i'm going to set up a VM and see if i can replicate their complaint [18:14] but i'm not 100% sure what 'file' they created in policy-rc.d [18:14] sdeziel: wanted to get your POV before I start deep diving :p [18:14] because this sounds like user has broken something xD [18:15] `/etc/nginx/nginx.conf.dpkg-new` is the only config I find when using a `exit 0` policy [18:16] it SOUNDS like whatever is being done in policy-rc.d is not changing the .dpkg-new config to a normal nginx.conf when nginx.conf doesn't exist [18:16] but that sounds more like a policy-rc.d problem than nginx [18:16] teward: for a reproducer: `lxc launch ubuntu-minimal-daily:24.04 n1` then `lxc exec n1 -- sh -c 'apt-get update -qq && echo exit 0 > /usr/sbin/policy-rc.d && chmod +x /usr/sbin/policy-rc.d && apt-get install -y nginx'` [18:16] *sips coffee* [18:16] sdeziel: stuck on a Windows system right now so gonna just spin from the 24.04 ISO in VMware [18:17] i'm at work so i don't have my default linux build >.> [18:17] yeah, that's also my understanding so far hence me asking why `exit 0` was used instead of the more common (to me at least) `exit 101` [18:18] sdeziel: it looks like they may be going off 2013 spec / documentation [18:18] so it's quite possible things changed since then. They're quoting Debian specs, that may not apply to Ubuntu [18:18] not sure if you saw their latest reply on that one [18:19] but figured before I deep dive into this being an nginx problem I was going to poke to get your POV [18:19] (and forgive me for setting it Incomplete, that's just me triaging xD) [18:21] teward: I just found something interesting, with `0`, it fails at `invoke-rc.d --quiet nginx status` during postinst [18:21] which feels like user-inflicted issue :) [18:21] that sounds like an invoke-rc.d bug then [18:21] sdeziel: the vast majority of bugs I see filed are problems with other packages or user configurations on end systems [18:22] so this is either invoke-rc.d being broke or something else, but I don't think it's an nginx bug or we'd see it in a lot of other cases. [18:22] teward: in that case, it seems the failure to be that `invoke-rc.d --quiet nginx status` tries to call the systemd service which then fails due to the `ExecStartPre` trying to validate the config while it's not there yet [18:23] sdeziel: this sounds like the debian bug i mentioend 1 momnet [18:23] ffs i can't type [18:23] 1 moment while I dig [18:23] #LotsOfEmail [18:23] was this under 24.04? [18:24] yes it is [18:24] 24.04, yes [18:24] wish people used ubuntu-bug to get the apport hooks triggering too [18:24] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1082373 [18:24] -ubottu:#ubuntu-server- Debian bug 1082373 in nginx "nginx: nginx missing pre-depends on nginx-common" [Important, Open] [18:24] this is new from today, suggesting 1.26.0 doesn't predepend on nginx-common and *can* trigger a case where the config doesn't exist [18:24] i think in Debian this is a side effect of the change to the package [18:25] but it may be related. [20:09] Hi running Ubuntu Server LTS 22.04.05 - my networking suddenly stopped working in the middle of the night last night - not sure of the time. I cannot connect to the Internet and networking software seems to have disappeared. Here's more: https://pastebin.com/QPvRTZ1t [20:09] Thank you in advance. [20:15] gmachine24: ubuntu server (unlike desktop) defaults to netplan and systemd-networkd for network management [20:17] you can also use the "ip" command to configure and bring uo network interfaces [20:19] Any idea how/why this happens? [20:21] Also, as an aside, I'm being offered the upgrade to the latest version of ubuntu server - I don't normally rush into doing this. [20:22] network interfaces don't go down by themselves normally. you will need to have system level access to review the logs to find out what happened. [20:23] Of course. Well - something changed and I again have network connectivity. [20:23] But I'll check the logs. [20:23] Thank you for your time. [20:23] in the best case it was just an outage at your remote servers' network connectivity provider [20:25] which http://status.THATORGANISATION.TLD would possibly be telling you about [20:25] I was just thinking something like that - it could have been a brief power outage. I live in a place with many electrical storms and occasional short-lived surges and grid failures. [20:27] that's what montiroing services are good for - at least you'll know what went down [20:27] network/service availability monitoring, i mean [20:28] yeah [20:29] it's so frustrating when you wake up and you think a job will be done and not only is the job not finished but there are more problems === utkarsh18 is now known as utkarsh2102