| === maclin1 is now known as maclin | ||
| ejat | if i want to install libcurl4 , The following packages will be REMOVED: | 02:58 |
|---|---|---|
| ejat | cryfs dotnet-runtime-2.0.0 dotnet-sdk-2.0.0 libcurl3 powershell slack-desktop virtualbox-5.2 | 02:58 |
| Unit193 | ejat: cryfs is in -proposed, others are external packages so complain to their respective sources? Or just wait for them to update when bionic is released. | 04:01 |
| Unit193 | And specifically, virtualbox's repo tends to lag a bit, so expect it a little after release. | 04:25 |
| === tacocat` is now known as tacocat | ||
| cpaelzer | bdrung: I created bug 1753938 for us to work on it | 08:06 |
| ubottu | bug 1753938 in qemu (Ubuntu) "slirp: domainname and classic stateless for DHCP" [Undecided,New] https://launchpad.net/bugs/1753938 | 08:06 |
| bdrung_work | cpaelzer, i haven't opened a bug report, but instead developed the patches and send them to the mailing list. | 09:09 |
| bdrung_work | cpaelzer, looking at #1753938, I missed the feature freeze with my upload | 09:10 |
| smoser | infinity if you're awake, we'd like to have rharper added to https://launchpad.net/~ubuntu-release-nominators/+members#active . | 09:11 |
| smoser | looks like you or bdmurray can do that , but i'm sure he's sleeping now. | 09:11 |
| smoser | infinity: sorry, 'raharper' is who needs adding. not rharper. | 09:13 |
| infinity | smoser: Done. | 09:15 |
| smoser | thank you | 09:18 |
| cpaelzer | bdrung_work: yeah, I think I dumped all we need to do on the bug | 09:21 |
| cpaelzer | let me know how you want to proceed once you are ready | 09:22 |
| cpaelzer | I assume you go on upstreaming anyway | 09:22 |
| cpaelzer | super-worst-case we can still revert the change in a few weeks, but I very much would prefer to clean up things and keep it | 09:22 |
| cpaelzer | bdrung_work: up to you | 09:22 |
| pitti | cpaelzer: guten Morgen, wie gehts? | 09:28 |
| pitti | cpaelzer: https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1696471/comments/9 "without a clear repro" → you are saying that my reproducer doesn't work? | 09:29 |
| ubottu | Launchpad bug 1696471 in libvirt (Ubuntu Xenial) "AppArmor denies access to /etc/gss/mech.d/" [Undecided,Confirmed] | 09:29 |
| bdrung_work | cpaelzer, i like to get the patches into upstream and update the package to use the accepted version of the patches | 09:49 |
| cpaelzer | pitti: I might have forgotten you had one but just found my own :-/ | 09:50 |
| cpaelzer | pitti: anyway I can drive it forward now | 09:50 |
| pitti | cpaelzer: I posted it to the bug a few days ago | 09:51 |
| cpaelzer | bdrung_work: ok, lets do that, if you have an accepted version get in touch with me | 09:51 |
| cpaelzer | bdrung_work: ok for yu? | 09:51 |
| pitti | but indeed it wasn't very hard, the xml isn't particularly magic | 09:51 |
| cpaelzer | you | 09:51 |
| pitti | cpaelzer: danke | 09:51 |
| cpaelzer | pitti: oh no, that is not it | 09:51 |
| cpaelzer | pitti: I can not trigger with your xml | 09:51 |
| cpaelzer | pitti: the magic ingredient is libsasl2-modules-gssapi-mit being installed | 09:51 |
| cpaelzer | it seems to be loaded and by that makes qemu read the dir | 09:52 |
| cpaelzer | without it won't | 09:52 |
| cpaelzer | which explains why I only saw it sometimes | 09:52 |
| pitti | cpaelzer: ah! I suppose that gets pulled in by a package like virtinst, sssd-tools, or perhaps realmd | 09:53 |
| pitti | I don't directly install any *sasl* or *gss* package | 09:53 |
| sarnold | cpaelzer: perhaps abstractions/kerberosclient needs to be updated, and perhaps something in libvirt needs to use this abstraction? | 09:53 |
| pitti | cpaelzer: nice, thanks for finding that | 09:53 |
| bdrung_work | cpaelzer, yes | 10:40 |
| xevious | rbasak: Last week, you mentioned that the mysql package changed how the service is started and stopped, which pointed me in the right direction to fix the problem I was having installing it in a `systemd-nspawn` container. | 16:07 |
| xevious | rbasak: When you install the mysql-server package, it used to enable the mysql systemd service. It doesn't get enabled automatically now, when I install it in a deboostrapped container. Is that an intentional change? | 16:08 |
| sarnold | xevious: at least the debian postinst appears to use invoke-rc.d; does your container have the policy in place that would prevent statup as a result? | 16:13 |
| Son_Goku | hey, was there a recent merge of updates to the compiler? | 16:16 |
| Son_Goku | it seems to have broken my ability to build qemu locally | 16:16 |
| Son_Goku | (as in, the qemu package that's currently shipped in bionic) | 16:16 |
| xevious | sarnold: The postinst started failing in my `systemd-nspawn` container, which I was able to fix by adding the `--as-pid2` parameter to `systemd-nspawn`, but maybe that broken enabling the service... | 16:17 |
| xevious | (mysql-server's postinst) | 16:17 |
| xevious | s/broken/broke/ | 16:18 |
| nacc | Son_Goku: which qemu are you trying to build? | 16:18 |
| Son_Goku | 1:2.11+dfsg-1ubuntu2 | 16:18 |
| Son_Goku | basically the top of the ubuntu-bionic branch of the qemu code | 16:19 |
| rbasak | xevious: the reason for the change was to fix https://bugs.launchpad.net/ubuntu/+source/mysql-5.7/+bug/1592669 | 16:19 |
| ubottu | Launchpad bug 1592669 in mysql-5.7 (Ubuntu) "postinst fails when daemon is not running (or is disabled by policy-rc.d)" [High,Fix released] | 16:19 |
| nacc | Son_Goku: ok, because i know ubuntu3 FTBFS | 16:19 |
| nacc | Son_Goku: there's an ubuntu4 in bionic-proposed | 16:19 |
| rbasak | xevious: that changes quite a bit about how it all works. The postinst has to start the daemon to upgrade the db, which we consider to be outside the scope of policy-rc.d. | 16:19 |
| nacc | Son_Goku: which did build, so it seems unlikely that the compiler is broken :) | 16:19 |
| Son_Goku | nacc, thanks, I'll take a look at that | 16:20 |
| rbasak | xevious: I think it's likely that stuff you're doing is colliding with that, because that's a new pid to any management thing like systemd. | 16:20 |
| nacc | Son_Goku: ah and ubuntu4 has a fix for FTBFS on newer glibc :) | 16:20 |
| nacc | Son_Goku: so yeah, just pull that one down, I think | 16:20 |
| Son_Goku | yeah, that looks about right | 16:20 |
| xevious | rbasak: Remember my other ticket, though? postinst failed without the `--as-pid2` parameter. | 16:20 |
| rbasak | xevious: I believe that the behaviour we have now is correct by policy though, so if you pinpoint something that shows it isn't, please let me know :) | 16:20 |
| rbasak | Yep | 16:20 |
| Son_Goku | nacc, we're bombing out on memfd_create() :) | 16:21 |
| Son_Goku | so that made me suspicious of compiler or other low level packages | 16:21 |
| rbasak | Because the postinst does actually start a mysqld instance, which expects to be a daemon and does daemon-like things. | 16:21 |
| xevious | rbasak: So, I'll need to manually enable mysql in my build process, since postinst won't be able to do any rc.d stuff inside my container, right? | 16:21 |
| rbasak | But that instance is only used within the postinst and is separate from the daemon instance started for you that provides normal service. | 16:21 |
| xevious | Yup | 16:22 |
| rbasak | xevious: I think there's still something wrong with your build process. | 16:22 |
| xevious | It's reproducible by deboostrapping a fresh copy of bionic and installing mysql with `systemd-nspawn`. | 16:22 |
| xevious | What's wrong with that process? | 16:22 |
| xevious | (that's an honest question, not me being stubborn/challenging/rude/etc.) | 16:28 |
| Son_Goku | RUUUUUDE! | 16:30 |
| Son_Goku | nacc, as an aside, the d/control for qemu still points Vcs-Browser to the utopic brance | 16:31 |
| Son_Goku | *utopic branch | 16:31 |
| nacc | cpaelzer_: --^ | 16:43 |
| xevious | rbasak: Actually, I totally misspoke: it does enable the service if mysql-server is the only thing being installed. I'm currently paring down the difference between the minimal process that works and my imaging script. | 17:04 |
| xevious | The `--as-pid2` parameter is definitely required, though. | 17:08 |
| rbasak | xevious: --as-pid2 sounds like it's needed for any use of nspawn that isn't an actual init program. | 17:27 |
| rbasak | Which also suggests to me that you're misusing it somewhat, though I'm not familiar with it of course. | 17:27 |
| xevious | In my experience, you can either fully boot a container (with the -b parameter) or use *most* programs without requiring --as-pid2. | 17:28 |
| pgoetz | I installed 18.04 on a workstation with dual GTX 1070 cards, one connected to a 4K monitor and am having the following odd issue. The first person to log in is locked down to a 1024 display and settings doesn't seem to recognize the monitor. However, if user switching is used to log someone else in, that user gets the full 4K display. Did not yet test with a 3rd user, as that is already weird enough. | 19:44 |
| wxl | pgoetz: #ubuntu for support | 19:44 |
| pgoetz | OK, thought this might be an issue with the beta. | 19:45 |
| pgoetz | join #ubuntu | 19:45 |
| wxl | if it were, you'd probably discuss it at #ubuntu+1 | 19:45 |
| wxl | sorry i failed to see the 18.04. that would probably be the best place | 19:46 |
| pgoetz | thank you. | 19:46 |
| xevious | rbasak: I've tracked down the source of my issue. PEBCAK, go figure. | 19:46 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!