[00:23] <cjwatson> stgraber: that looks right, yes
[01:29] <antarus> cjwatson: stgraber curious how interested you are in automated installer error reporting
[01:30] <antarus> like the ability to preseed some sort of address to send reports ?
[01:56]  * cjwatson gets enough bug reports, personally ;-)
[01:57] <antarus> well we will likely prototype this internally
[01:57] <antarus> but I'd like to send teh patches upstream
[01:58] <antarus> because I don't want to maintain a d-i / u-i fork
[01:58] <antarus> ubuntu doesn't necessarily have to use the feature
[01:58] <antarus> but we would ;p
[01:58] <cjwatson> base it on the installation-report package I guess
[01:58] <cjwatson> I'm not opposed to carrying patches we don't use
[02:00] <antarus> hmm
[02:02]  * cjwatson glares at ... whatever made the ubiquity test suite start failing
[02:03] <cjwatson> well, guess I'll look into that tomorrow.  night
[02:03] <antarus> night
[02:04] <antarus> I'll poke at the report package
[02:18] <stgraber> cjwatson: cool, I'll upload a new netcfg with that fix and a d-i tomorrow morning unless you beat me to it or have something else you want in there
[02:21] <antarus> stgraber: did you push an initrd with the resolvconf fixes?
[02:22] <stgraber> antarus: now, that'll happen with the next d-i upload. It'll take a while for netcfg to build and publish and I likely won't be around at that time (east coast), so I'll upload a new debian-installer to pick up these changes tomorrow morning unless cjwatson does it before
[02:22] <stgraber> *no
[02:23] <antarus> ok
[02:24] <stgraber> antarus: so assuming you're on the west coast, you should have new d-i images tomorrow morning
[02:24] <antarus> excellent
[02:24] <antarus> my co-worker *really wants* installs to work, so I have passed him the news ;)
[02:47] <stgraber> antarus: is that Thomas? :)
[03:05] <antarus> stgraber: yes
[03:05] <stgraber> antarus: hehe, ok, he also e-mailed me about it earlier ;)
[03:06] <antarus> stgraber: ahh sorry about that ;p
[03:06] <antarus> not trying to hound you
[03:07] <stgraber> that's fine, I'm actually really happy not to be the only user of these features and have people actually test the development release rather than filing bugs when we've already released (with the whole problem of getting a new installer in a release version of Ubuntu)
[03:07] <antarus> we aim to release our internal product like 2-4 weeks after you
[03:07] <antarus> this is probaly the earliest we have really started
[03:08] <antarus> usually we start much later and it goes badly ;)
[03:09]  * antarus is | | this close to running Precise as his primary desktop
[03:10] <stgraber> hehe, yeah, I've been there in a past life ... anyway, if you find anything weird network related, feel free to ping me, I really want that to be rock solid for 12.04 as fixing that stuff post-release is a pain (as I discovered when trying to fix some 10.04 bonding bugs ...)
[11:12] <cjwatson> stgraber: sounds good to me, thanks
[11:15] <cjwatson> stgraber: could you please push your netcfg changes to lp:~ubuntu-core-dev/netcfg/ubuntu, though?
[14:05] <stgraber> cjwatson: sure
[14:36] <stgraber> cjwatson: pushed everything to the branch which made me notice my ubuntu10 upload contained some extra debug statements that I didn't notice in the debdiff (because for some reason my previous .dsc also had said debug statements ...)
[14:36] <stgraber> so I just uploaded ubuntu11 without them and will upload d-i once it's published
[14:37] <stgraber> (doesn't really make any difference besides a few "X: got_lease = %d" everywhere we touch/check got_lease but we probably don't want that anyway ;))
[17:08] <bdmurray> stgraber: in bug 924538 I noticed the following
[17:08] <ubot2> Launchpad bug 924538 in ubiquity "Failed to install all packages selected in Edubuntu according to error popup" [Medium,New] https://launchpad.net/bugs/924538
[17:08] <bdmurray> Jan 31 20:39:36 edubuntu kernel: [ 1001.203724] EXT4-fs (sda1): mounted filesystem with ordered data mode. Opts: (null)
[17:08] <bdmurray> Jan 31 20:39:36 edubuntu kernel: [ 1001.864039] blockdev: sending ioctl 125d to a partition!
[17:08] <bdmurray> Jan 31 20:39:36 edubuntu kernel: [ 1001.864039] blockdev: sending ioctl 125d to a partition!
[17:09] <bdmurray> Jan 31 20:39:36 edubuntu kernel: [ 1001.893181] blockdev: sending ioctl 125d to a partition!
[17:09] <bdmurray> does that mean anything to you?
[17:15] <stgraber> I remember seeing this during some installs, I'm not exactly sure what's going on though.
[17:15] <stgraber> https://bugzilla.redhat.com/show_bug.cgi?format=multiple&id=783955 seems to point to --setro and just an informational message, not an error
[17:43] <bdmurray> stgraber: how does the hostname lookup in ubiquity work? bug 924511
[17:43] <ubot2> Launchpad bug 924511 in ubiquity "ubiquity tells me my computer name already exists on the network when it doesnt" [Medium,New] https://launchpad.net/bugs/924511
[17:46] <stgraber> I believe it does an avahi query, similar to pinging <hostname>.local
[17:46] <GrueMaster> I see that all the time in oem-config, but I assumed it was because all of my test systems are defined in my dhcp server with mac addresses, names, and fixed ip addresses.
[17:46] <GrueMaster> If that's the case, other systems may be caching avahi data.
[17:47] <GrueMaster> Ah, read the bug.  This is different.
[17:47] <stgraber> bdmurray: it's essentially doing "getent hosts <hostname>" and "getent hosts <hostname>.local"
[17:48] <stgraber> bdmurray: if one of the two returns something, you get the error
[17:50] <GrueMaster> I wonder if it is timing related in that it is testing with a blank field which should return a different error.
[17:51] <GrueMaster> stgraber: On a side note, has d-i been uploaded?  Can we expect a new netboot today?
[17:52] <stgraber> GrueMaster: not uploaded yet, waiting for netcfg to build on powerpc (taking a long long time apparently). Will upload d-i right after that, so should be ready later today, yes
[17:52] <GrueMaster> ok.  Kind of holding up automation here, but not critical (yet).  :P
[17:55] <stgraber> bdmurray, GrueMaster: That's basically what ubiquity does: http://paste.ubuntu.com/832913/
[17:55] <stgraber> that test takes a while so it's done using the _async function in ubiquity which means it can take up to 4-5s before the warning shows up or disappears
[18:27] <CIA-2> debian-installer: stgraber * r1623 ubuntu/debian/changelog: releasing version 20101020ubuntu105
[18:28] <stgraber> just waiting for the powerpc package to publish and will upload this one ^
[18:31] <stgraber> GrueMaster, antarus: new d-i uploaded
[18:32] <GrueMaster> excellent.  Will keep my eyes open for it.
[18:33] <bdmurray> stgraber: so the bug should be won't fix then since it just takes a bit?
[18:37] <stgraber> bdmurray: yeah, though looking at the comments it's not clear to me that the warning indeed disappeared after a while for him. Though IIRC the check is also done when changing field which explains why moving to username made the warning disappear
[18:38] <stgraber> bdmurray: we probably could improve that somehow by giving some kind of feedback that it's being checked
[18:39] <stgraber> bdmurray: so I'd keep the bug open as a generic "let's try to improve the hostname validation code" unless we have another one like that already, priority low and assigned to me would be fine I guess
[18:39] <bdmurray> stgraber: okay, low sounds better then
[18:41] <stgraber> gah, d-i failed to build because of a font package ... didn't think I'd have to do a test build for a no-change rebuild ...
[18:58] <CIA-2> debian-installer: stgraber * r1624 ubuntu/build/pkg-lists/gtk-common: gtk-common: Replace ttf-thai-tlwg-udeb by fonts-thai-tlwg-udeb
[18:58] <CIA-2> debian-installer: stgraber * r1625 ubuntu/debian/changelog: releasing version 20101020ubuntu106
[18:59] <stgraber> and uploaded
[19:48] <bdmurray> stgraber: regarding the hostname checking I've found a couple of other bugs and was wondering if you had any ideas about a tag to identify them.  Does it make sense and would you or somebody use it?
[19:57] <stgraber> bdmurray: are they similar bugs? I don't expect there'd be enough different bugs related to that single field that we need a tag for it, but well, you can always be surprised at how many bugs can hide behind 20 lines of code ;)
[19:58] <bdmurray> stgraber: I'm more curious about generally grouping types of ubiquity bugs together (location map, hostname, partioning etc) than just the hostname part
[19:59] <stgraber> it'd probably make sense to group them by step indeed
[20:00] <bdmurray> step being "Who are you?" in this case?
[20:00] <stgraber> ubi-console-setup, ubi-language, ubi-migrationassistant, ubi-network, ubi-partman, ubi-prepare, ubi-tasks, ubi-timezone, ubi-usersetup, ubi-warning.disabled, ubi-webcam, ubi-wireless
[20:01] <stgraber> well, you can probably ignore ubi-warning.disabled as it's, well, disabled ;)
[20:01] <bdmurray> stgraber: ah great!
[20:01] <stgraber> if we could use these as tags it'd probably be useful as it's a direct mapping to the code
[20:03] <bdmurray> right and maybe put screenshots of pages / with tags at debugging ubiquity
[20:36] <bdmurray> stgraber: I've run into the wget canont resolve archive.canonical.com bug when installing now
[20:36] <bdmurray> however trying a wget after ubiquity has already crashed worked for me
[20:36] <bdmurray> so I'm really not certain what when awry
[20:56] <stgraber> bdmurray: it's not impossible you actually had a DNS failure, though that'd be pretty bad luck. If you can reproduce the failure at the same place a few times in a row, then it might be something odd going on with resolvconf/dnsmasq
[20:58] <bdmurray> stgraber: based off the comments in bug 870643 (which a bugpattern points / pointed to) I'm inclined to think the latter
[20:58] <ubot2> Launchpad bug 870643 in flashplugin-nonfree "package flashplugin-downloader failed to install/upgrade: wget: unable to resolve host address `archive.canonical.com'" [High,Confirmed] https://launchpad.net/bugs/870643
[20:59] <stgraber> bdmurray: oh, actually, it's pretty likely to be the exact same issue I fixed yesterday in d-i/netcfg
[21:00] <stgraber> bdmurray: hmm, or not, the exact same issue can't appear in ubiquity, but something similar could ...
[21:00] <stgraber> bdmurray: can you get /proc/mounts, /target/run/resolvconf (all of it) and /run/resolvconf at the moment of the crash?
[21:01] <bdmurray> stgraber: at the moment of the crash? the installation continues even though flashplugin-installer fails
[21:03] <stgraber> bdmurray: yeah, that's the tricky part, you'd probably have to watch /var/log/syslog for the error to appear, though you can't use tail because inotify is broken on overlayfs ...