=== StevenK_ is now known as StevenK === kentb_ is now known as kentb-out === artnay_ is now known as artnay === ara is now known as Guest69549 [12:55] Are there any books out which detail use of cobbler (orchestra) in the datacenter for large-scale system deployment? [13:11] fixxxermet: entirely wrong channel for that question. Try #ubuntu-server [13:12] ev: thanks [13:12] sure thing! [13:12] good luck [13:13] hmm, my boss is idling in there. [13:13] if you dont get along with the installer chapter in that book and find a bug, please come back though :) [13:13] good morning [13:27] ev: hey! While working on the gsettings changes yesterday, I noticed that prepare() in apt_setup doesn't seem to be called/used during install, at least looking at all the gsettings keys being accessed I couldn't find any access to the HTTP proxy keys [13:28] ev: do you know if that part of the code is supposed to work/be used and what exactly uses it so I can test that the HTTP proxy stuff actually works [13:31] .prepare should be called from FilteredCommand [13:32] indeed [13:32] I would be astonished if it weren't called, because it's the thing that defines what command is called [13:32] if it weren't called then it wouldn't work at all [13:33] check that there aren't import failure exceptions or something? [13:33] I'll be doing some more tests once I'm done zsyncing today's daily and add some debug to both the gsettings module and apt_setup.py itself [13:36] usb-creator: cjwatson * r359 trunk/debian/ (changelog rules): [13:36] usb-creator: Build with dh_translations, to localize the .policy file at runtime. [13:36] usb-creator: LP: #853227 [13:38] usb-creator: cjwatson * r360 trunk/debian/ (changelog control): [13:38] usb-creator: Build-depend on dh-translations and bump required debhelper version, to [13:38] usb-creator: support previous change. [13:38] thanks for that [13:40] ubiquity: stgraber * r5032 ubiquity/debian/changelog: Update changelog [13:43] usb-creator: cjwatson * r361 trunk/debian/changelog: releasing version 0.2.34 === kentb-out is now known as kentb [13:55] ev, cjwatson: Just did an install using today's daily with apt_setup.py set to print something at the beginning of prepare() and at the beginning of the gsettings function + patched gsettings to print every key that's read or modified. I see all of the usual keys being accessed/changes but no reference to the proxy stuff... [13:56] and no import/stacktrace errors in the debug log [13:59] what is mirror/http/proxy set to? [14:00] oh, interesting, it's actually set to http://http://172.21.20.1:3128 which seems to indicate that the code worked (though logging didn't) [14:00] well "worked", that value is invalid but at least it's set to something :) [14:02] ubiquity: stgraber * r5033 ubiquity/ (debian/changelog ubiquity/components/apt_setup.py): Fix PROXY url as the gsettings host key already contains http:// [14:03] is it guaranteed to contain http:// ? [14:03] I wonder if it would be worth e.g. adding http:// only if it didn't contain a colon, or something [14:04] I guess it'd be safer indeed. It seems to always start with http:// on my systems, but there's no clear specification on it that I could find. [14:07] cjwatson: I'm going to use .startswith("http://") and if it doesn't, then prepend it. This should take care of the case where we get an IPv6 [14:07] * cjwatson nods [14:07] (should've thought of that :-) ) [14:08] ubiquity: stgraber * r5034 ubiquity/ubiquity/components/apt_setup.py: Make sure that the proxy host starts with http://, if it doesn't then prepend it [14:09] now to finally fix bug 657086 :) [14:09] Launchpad bug 657086 in ubiquity "No restart option after choosing to "Continue Testing"" [High,Won't fix] https://launchpad.net/bugs/657086 [14:10] ev: stealing that one from you ^ [14:12] by all means [14:12] :) [14:12] trying to finish off that resize considering swap set of patches, keep getting pulled into other things :-/ [14:44] ubiquity: stgraber * r5035 ubiquity/ (debian/changelog ubiquity/frontend/gtk_ui.py): Fix gtk_ui to only set the gsettings keys once as for some reason these functions seem to be called multiple times... [14:45] and that's bug 657086 fixed (at least it disappeared here :)) [14:45] Launchpad bug 657086 in ubiquity "No restart option after choosing to "Continue Testing"" [High,Won't fix] https://launchpad.net/bugs/657086 [14:45] yay [14:46] still wondering why these functions get called multiple times though, that'd be the real bug though making sure a key can only be changed once is still a good thing to check... [14:54] ev: looking at bug 651932. Should we also make the "Install this third-party software" option depend on network being available? (looking at the code, this seems to install ubuntu-restricted-addons which isn't on the CD) [14:54] Launchpad bug 651932 in ubuntu-release-notes "Without internet connection, the installer shouldn't recommend the updates during installation, or third party software" [Undecided,Fix released] https://launchpad.net/bugs/651932 [14:59] stgraber, doesn't it also call jockey to install wireless drivers (which are in the pool)? [15:02] ev: I noticed if I'm connected to wired and try to connect to wireless in the installer, I can never click "Continue", the button remains "Connect". I'm under the impression that's because the code is checking for what the default connection is, rather than just what's active for a particular device [15:02] is this already reported as a bug? [15:03] cyphermox: not to my knowledge. Please file one and assign it to me, providing as much details on the steps needed to reproduce as possible [15:03] sure [15:03] do make sure you're using the latest version too. Newer versions of ubiquity should skip the wireless page entirely if you're on wired ethernet [15:03] assuming you can see the Interwebs [15:04] superm1: it could be, but then the description is going to be inaccurate on systems that run without internet connectivity (as the description doesn't mention drivers but only MP3 support) [15:04] ev: then i'm most likely not using the latest version; I was using 20111004 livecd. [15:05] mm, that should have it [15:05] perhaps there's a bug there [15:05] cyphermox: that should be ubiquity 2.8.2 [15:05] yu[ [15:05] *yup [15:06] ok, I didn't commit anything related to network/wireless after that release, so any network/wireless problem you see will also happen with trunk [15:07] in the live session I can't seem to get back to the wireless page, just a second [15:09] and I no longer get to the wireless page, as expected [15:11] superm1: ok, that was just me not reading properly ;) so indeed the UI says it's both drivers + addons. Drivers are on the CD, addons aren't. I'll quickly make sure we can still install with that option selected when internet isn't available. [15:12] superm1: if that's the case I'll just mention it in the bug and close it as the download updates part has been fixed [15:12] stgraber: seems it's just because NM might be slow to establish the connection. If the connection is established to wired before the welcome dialog pops up, the wireless page is skipped, otherwise you'll get to it [15:19] * cjwatson tries to disentangle LANG vs. LC_* logic from language-selector [15:19] y'all should speak English [15:20] Is that British English or English? [15:21] yes [15:21] lol [15:23] well played :) [16:28] Xubuntu has a problem with ubiquity crashing on today's desktop image - bug 867620 [16:28] Launchpad bug 867620 in ubiquity "Xubuntu failed to install, Ubiquity crashed" [Undecided,New] https://launchpad.net/bugs/867620 [16:29] I can confirm that is crashes a few minutes after entering name/password/etc [16:33] Do you need a ubiquity --debug for this? [16:34] stgraber: ^- that looks like yours [16:35] okay, I've created https://code.launchpad.net/~ev/partman-partitioning/save-swap-size-real/+merge/78122 and updated https://code.launchpad.net/~ev/ubiquity/factor-swap-in-resize/+merge/77973 for the resize fix [16:35] If it helps, the Xubuntu desktop images shrunk last night by about 9MB [16:36] if I could have another pair of eyes on that before I merge, I'd greatly appreciate it [16:38] cjwatson: looking [16:39] also, if anyone has ideas for a fun 5 minute talk on some aspect of automated testing, do sign up: https://wiki.ubuntu.com/UDS-P/TestingInUbuntu [16:39] trying to foster knowledge sharing at UDS [16:40] charlie-tca: I'll have a fix in the next ubiquity, do you want me to paste you a workaround too? [16:41] no, we can wait. Will that be in tomorrows images? [16:41] I think so, we'll probably have enough changes in ubiquity for a release later today [16:41] Great! Thank you [16:42] If we are looking at Friday to start smoke testing the final images, we need to make sure they work [16:48] Does this crash have anything to do with http://cdimages.ubuntu.com/xubuntu/daily-live/ being our of sync? [16:49] Xubuntu current is dated 2011-10-04, but the listings only show 2011-10-03 and 2011-10-02 [16:49] Never mind; browser refresh fixes that [16:51] ubiquity crashes wouldn't affect CD builds anyway, no [16:52] Thank you [16:52] ubiquity: stgraber * r5036 ubiquity/ (debian/changelog ubiquity/gsettings.py): Better handle the case where a gsettings key isn't set. [16:53] ubiquity: stgraber * r5037 ubiquity/ (debian/changelog ubiquity/components/apt_setup.py): Don't crash when the PROXY ignore-hosts list is empty [16:53] charlie-tca: ^ there you go [16:55] charlie-tca: I'll do a quick test install of xubuntu to make sure we don't get some other gsettings weirdness as XFCE uses gsettings but doesn't seem to completely share the schema with gnome [16:55] Okay, thank you [17:04] I'll start poking at bug 813065 while Xubuntu downloads [17:04] Launchpad bug 813065 in casper "Live session switches to VT console briefly" [High,Confirmed] https://launchpad.net/bugs/813065 [18:31] ubiquity: stgraber * r5038 ubiquity/ (bin/ubiquity-dm debian/changelog): Clear the console just before killing X in ubiquity-dm, this should give us a blank screen while waiting for lightdm to start. [19:59] can the ubiquity task in bug 855042 be closed? [19:59] Launchpad bug 855042 in ubiquity "ubiquity hangs while jockey-text at 100% CPU" [High,Confirmed] https://launchpad.net/bugs/855042 [20:10] charlie-tca: good thing I tested or tomorrow's Xubuntu build would have failed just a line after today's ;) [20:11] Wow! [20:11] This is not a good time for those to fail, either. [20:12] yeah, just making sure I can install xubuntu i386 + boot it, then I'll commit what I have so tomorrow's ubiquity at least lets you install :) [20:12] Thank you for being persistent [20:12] I'll also do a quick test install of kubuntu just to make sure I didn't mess up anything there with my gsettings changes [20:17] ubiquity: stgraber * r5039 ubiquity/ubiquity/components/apt_setup.py: One more fix to apt_setup to skip proxy config also if gsettings returns None [20:18] charlie-tca: xubuntu installs with current trunk (at least on i386) [20:19] That works for me [20:20] stgraber: you do know I will whine tomorrow if amd64 fails, right ;) [20:20] I suspect if i386 works, amd64 will also, though. [20:20] I'm sure you will :) [20:20] though all the changes I did applied to the python code, so pretty unlikely to be arch specific :) [20:21] That was my thought too. [20:44] cjwatson, ev, infinity: Ok, so current ubiquity trunk fixes quite a few bugs and has been tested on ubuntu, xubuntu and kubuntu. Any of you planning on pushing changes in the next hour or can I upload what we have? [20:45] I'd like to have it uploaded soonish so it makes it to the next Edubuntu daily build (one of the first ones to build apparently) [20:56] ubiquity: stgraber * r5040 ubiquity/bin/ubiquity-dm: Update comment in ubiquity-dm (s/upstart messages/boot-time messages/g) [22:08] stgraber: go ahead, I still need to fix up some locale handling but I'm not going to do that tonight [22:10] stgraber: subprocess.Popen(['clear']) - shouldn't we wait for that to exit to avoid a race condition, i.e. subprocess.call instead?