veebullwhere can a person find some details about what options do what in the Server installer?07:17
veebullI've looked thru the Server Guide and dug around in the Community Docs and not finding what I'm looking for07:17
holsteinhey veebull07:24
holsteinwhat are you looking for?07:24
veebullfor starters... when I boot a 32bit i386 Server install CD, after choosing the language the screen with the F1-6 options at the bottom...07:25
holsteini would just google each one07:26
holsteinive really only messed with a few07:26
veebullWhere can I find an explanation of those options?  E.g. F4 'minimal system', 'minimal virtual machine', 'oem install', etc.07:27
veebullI've dug and dug, and I'm slowly finding bits and pieces here and there - I guess the 'OEM' install is a takeoff of the Kubuntu oem install stuff, but there appears to be next to no current info on it07:27
veebullthe minimal virtual machine appears to be related to some sort of JeOS install (stripped down and lots of drivers removed as un-necessary for a virtual machine environment, plus some optimizations - but again nothing in the official documents07:28
holsteini bit here about that http://blogbuildingu.com/articles/ubuntu-installation-guide07:30
holsteinveebull: i wouldnt try to figure out all the options07:30
holsteintheres plenty ;)07:30
veebullam I expecting too much that installer options presented to the user should be in the 'official' documents like the Server guide, not hidden away on a wiki page that says 'yeah we need to get around to doing this some day'?07:30
holsteini would just try and get the right install for your case07:31
veebullI know... I'm just remembering the bad old days of SuSE's gargantuan 'tome' of a user manual ;)07:31
holsteini learned alot reading about and trying some of thses07:31
veebullAlso...  i picked the F4 'minimal install', and towards the latter portion up pops the tasksel screen... with options like none I've ran into before07:32
holsteinpretty sure they're all lucid now too :)07:32
holsteinveebull: OH07:32
holsteinlike ssh server and LAMP or whatever?07:32
veebullWhat exactly is a 'Virtual Machine Host' on that menu?  No info, no description, not mentioned anywhere07:32
holsteinwell, i think its going to be just that07:33
veebullSame sort of menu, but with a few little twists thrown in, and no description of what differentiates each one07:33
holsteina server on which to host vitrual machines07:33
veebullDoes it have KVM, XEN, QEMU...?07:33
veebullWhat makes it special from a Basic Ubuntu Server install?07:33
holsteini suppose if you want it to07:33
holsteini was checking out http://www.howtoforge.com/vboxheadless-running-virtual-machines-with-virtualbox-2.0-on-a-headless-ubuntu-8.04-server07:34
holsteinand i had assumed that would give you an option like that07:34
holsteinfor xen maybe07:35
holsteinbut again, ive never needed to find out yet ;)07:35
veebullthats my point... for server stuff, should we have to be guessing what the install options do?07:36
holsteini think there is a bit of the idea that if you need it, you know what it is07:36
holsteinand you probably dont need any explanation07:36
holsteinbut there's going to be man pages somewhere07:37
veebullTrue... but man pages are only useful *after* you install the system... and find out that you didn't get what you wanted, and have to *re* install :(07:37
holsteini try and do several test installs anyways07:38
holsteinin vbox or something07:38
veebullAnd as far as 'if I need it, I know what it is'... I might know what I would want, but whether thats what the developer(s) were thinking... who the heck knows07:38
veebullSo we're back to hunt-n-peck and guess and try and guess again...07:39
holsteinwell, you can send an email07:39
veebullI'm not against learning by doing and exploring... but it seems like the installer options are getting changed faster than the documentation for them07:39
holsteinand the wiki's are pulic editable07:39
holsteinso, in theory, you can research it, and report :)07:40
holsteinveebull: and that could be the deal07:40
holsteinthe options have changed quite a bit from hardy to lucid07:40
veebullemail to who?  I've asked ubuntu.stackexchange.com, ubuntu forums, ubuntu-users list, ubuntu-servers list, irc ubuntu/ubuntu-server...07:40
holsteinand maybe thats just where he docs are right now07:40
holsteina little behind07:41
holsteinveebull: the ubuntu sever list would be the one07:41
veebulllooks like the LiveCD is somewhat documented under boot parameters  in the community docs, but not the Server CD07:41
holsteinveebull: has it been a couple weeks?07:41
holsteinwithout a response?07:41
holsteinveebull: the liveCD is for a whole other target audience though07:42
veebull3-5 days with zero response07:42
holsteinveebull: yeah, you might want to wait a month07:42
veebullyou're kidding, right?07:42
holsteinabout waiting a month07:43
holsteinwe dont know who is on that team07:43
holsteinand if they get paid07:43
holsteinif its a hobby07:43
holsteinif they are slammed07:43
holsteini say a month is generous07:43
holsteini mean07:43
holsteinthere's a couple ways of looking at this07:43
holstein1. theres no documentation for the server CD, what gives?07:44
holstein2. what can i do to improve the sever cd documentation07:44
holsteinand that might be more than you want to take on07:44
holsteinmaybe a strongly worded email is enough07:44
holsteinheres the team page07:45
holsteinlooks pretty active07:46
veebullWell, I'll have to give that a try as well.  I may need to wait a day lest I open mouth, insert foot and chew vigorously as I'm a bit miffed about the whole deal right now07:47
holsteindo what you feel is appropriate07:47
holsteinand i tottally agree07:47
holsteinsometimes you have to hang in the IRC channels for a couple days to though ;)07:47
veebullI need to find a compelling argument to squash the 'if you need it you should already know it' attitude as I have ran into that once or twice already in this search...07:49
holsteini think its because its all command line07:49
holsteinand a certain skill set is implied07:50
holsteinand an admin backgroud too maybe07:50
holsteinshould be documented07:50
holsteinveebull: check out https://wiki.ubuntu.com/ServerTeam07:51
holsteinNEXT MEETING: Tuesday 2010-09-28 at 1800 UTC - #ubuntu-meeting07:52
holsteini would think you could find some folk there to discuss these issues07:52
holsteinmaybe hang til the end07:52
holsteinand catch someone in a query or something07:52
veebullWorth sitting in or recording at least to get a feel for the personalities, at least.07:53
veebullWell, I better git.  Bedtime for bozo and all that.07:53
ScottKveebull: Particularly if you catch sommer on #ubuntu-server, he coordinates most of the Ubuntu Server guide development.  I'm sure he'd be glad to include stuff like this if someone would put something together.07:53
evand now to find out why 'ready when you are' isn't getting translated11:04
evincidentally, ntfsclone + nc = win11:05
CIA-28ubiquity: evand * r4397 trunk/ (debian/changelog scripts/simple-plugins):11:52
CIA-28ubiquity: Don't let Jockey's automatic driver installation failing cause the11:52
CIA-28ubiquity: entire prepare page to fail.11:52
evmat_t: ^ thanks for running into that :)11:52
cjwatsonogra: can you review https://code.launchpad.net/~mwaddel/ubuntu/maverick/flash-kernel/vexpress-support/+merge/36472 ?12:34
* ogra looks 12:34
ogracjwatson, looks ok, do you know if we have a bug for the changelog ?12:36
ograah, ignore me, i'm blind12:37
cjwatsonif you could deal with merging it, that'd be good12:38
ograon my way12:39
ograhmm, do we have any preseed option to avoid showing the "encrypt my homedir" checkbox ?13:11
* ogra wonders what else to do with bug 64642113:12
ubot2Launchpad bug 646421 in ubiquity (Ubuntu Maverick) (and 1 other project) "oem-config process that zero's swap broken on armel preinstalled images (affects: 1) (heat: 8)" [High,New] https://launchpad.net/bugs/64642113:12
persiaogra, Unless ubiquity is in automatic mode, the preseeding just prepopulates the dialogs, so users can change the preseed values.13:19
ograouch, ok13:19
* ogra knew that, just didnt remember it13:19
persiaProbably better to hunt down the bit in user-setup-apply that does that, and have it verify it's writing to a block device: if writing to a file, it should limit itself to the current size of the file.13:20
ograyes, sadly nothing i can tackle now (i'm in meetings from the full hour on until EOD) ... i was hoping for a quick frix through a preseed13:21
persiaI don't believe that to be possible, given the architecture.  I may be mistaken.13:22
ograno, indeed13:22
ograand indeed its not executed by ubiquity but by d-i13:25
cjwatsonit would be possible to add something preseedable in ubiquity to suppress the option13:27
cjwatsonnot there right now though13:27
persiaWell, err, that gets complicated.  ubiquity (as oem-config in this instance) runs the controller, which interacts with the d-i component, but since at that point the d-i component is bundled as part of oem-config...13:27
ograah, i found the function13:28
ograhmm, having GrueMasters /etc/crypttab would be helpful here13:29
ograbut i bet he wiped the install after filing the bug13:30
ogra                                if swapoff $device; then13:31
ogra                                        dd if=/dev/zero of=$device 2>/dev/null || true13:31
ogra                                fi13:31
ograthats the problematic part13:31
ograi wonder if we can safely assume that if $device doesnt start with /dev (or UUID=) it will be a file ...13:32
ogradoesnt really feel safe though13:32
persiaRight.  Needs to test whether $device is a block device or a file (if [ -b "$device" ] ...)13:32
ograoh, indeed13:32
* persia doesn't believe dd can handle UUID= anyway13:33
ograand then just do a stat13:33
ograto get the size13:33
persiaIn the [ -f ... bit yeah.13:33
persiaand if -b and -f both fail, whine about not being able to zero the swap and get on with life.13:33
ograthats a security hole though13:34
persiaThe whine loudly through debconf, and add a warning handler to the oem-config controller.13:34
persiaIt's a state machine: you can go back, and decide not to encrypt the home directory.  Just talk to the user about it.13:35
persiacjwatson, Just to make sure I understand your comment: you're saying that it would be possible to invent a preseed *to ubiquity* that would cause ubiquity not to display the UI component, rather than that preseeding some value of the selected choice would override ubiquiity's representation of the user selection to user-setup?13:40
cjwatsonpersia: sure, perfectly doable13:47
cjwatsonhaven't remotely thought about it to decide whether it's sane, since I'm not really doing much ubiquity stuff this cycle13:47
cjwatsonbut ubiquity is entirely capable of talking to debconf itself13:47
ev(see ubiquity/install_bootloader for one such example of it hiding UI based on a debconf preseed)13:48
ograwell, i guess its the same amount of work for either of the possible fixes13:50
persiacjwatson, Thanks for the confirmation.13:50
ograso i'd rather go with adding support for swapfiles to user-setup-apply13:50
mat_tev: totally involuntarily :)14:03
CIA-28ubiquity: evand * r4398 trunk/ (3 files in 2 dirs):14:23
CIA-28ubiquity: Make sure $LANGUAGE gets set in the parallel debconf child process,14:23
CIA-28ubiquity: so that we get translated descriptions (LP: #646109).14:23
evit occurs to me that if ubi-timezone changes the locale, it wont be reflected in the parallel plugins; however, this strikes me as an intractable problem as a parent process cannot change the environment of its child.14:25
=== ara__ is now known as ara
cjwatsonev: is it a bug if a town shows up multiple times in the list that appears if you type at the timezone page?14:32
cjwatsonactually, heh, I bet I know why this is happening - we aren't showing enough detail14:32
cjwatsonI typed Cambridge and got "Cambridge (England, United Kingdom)" or similar twice14:32
cjwatsonbut of course there are two Cambridges in England - the well-known one in Cambridgeshire, and the less-well-known one in Gloucestershire14:32
cjwatsonI wonder if the geonames data just doesn't distinguish ...14:33
ograand what about Cambridge near Boston ?14:33
cjwatsonnot in "England, United Kingdom"14:34
cjwatsonthere are plenty of other Cambridges shown14:34
ograoh, you already have the country info14:34
evcjwatson: well, it should definitely do the right thing.  Can you please file a bug with the details?14:36
evfor what it's worth, you can query the database at http://www.geonames.org/ and http://geoname-lookup.ubuntu.com/?query=cape%20town to see what it's actually producing vs what the real data is or what it's showing on the screen14:37
cjwatson{"name" : "Cambridge", "admin1" : "England", "country" : "United Kingdom", "longitude" : "0.116670", "latitude" : "52.200000" }, {"name" : "Cambridge", "admin1" : "England", "country" : "United Kingdom", "longitude" : "-2.366670", "latitude" : "51.733330" }14:38
Riddellev: has there been any changes to the privilages in ubiquity recently?  it's not starting on today's live CD  bug 64682714:53
ubot2Launchpad bug 646827 in ubiquity (Ubuntu) "ubiquity KDE frontend does not start (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/64682714:53
evyes, we call setgroups now14:53
cjwatsonbollocks, my fix to ubiquity's upstart job doesn't seem to have had any effect14:53
evRiddell: interesting, looking into it now14:54
cjwatsonnot getting plymouth on shutdown here14:54
evand scott is only reachable via 4square or something silly like that14:55
cjwatsonyeah, fuck that, I'll work it out myself14:55
cjwatson(actually e-mail works too)14:55
evI do believe that's one of the first times I've seen you properly curse in here14:55
cjwatsonI try to censor myself :)14:55
evI guess plymouth has that effect ;)14:56
cjwatsonyou might say that.  though I'm mostly being irritated right now by something I know Scott's irritated by as well14:56
cjwatsonnamely bug 554172, which I'm fairly sure is causing bug 54413914:56
ubot2Launchpad bug 554172 in upstart (Ubuntu Maverick) (and 5 other projects) "system services using "console output" not starting at boot (affects: 210) (dups: 17) (heat: 943)" [Medium,Fix released] https://launchpad.net/bugs/55417214:56
ubot2Launchpad bug 544139 in consolekit (Ubuntu Maverick) (and 3 other projects) "Active VT tracking can fail at startup (affects: 48) (dups: 6) (heat: 216)" [High,Triaged] https://launchpad.net/bugs/54413914:56
evUGH, thanks for thinking stdout was a file on disk, partimage14:57
evoh I think I saw this one, actually14:59
evI recall him talking about exposing a kernel bug in that you could not successfully reopen /dev/console, and having a bit of a spat with Andy over it15:00
CIA-28ubiquity: evand * r4399 trunk/ (d-i/update-control debian/changelog debian/control): depends on latest libindicator-dev for ABI change (LP: #637692)15:02
cjwatsonthat's the one15:04
rgreeninghey ev15:36
evrgreening: hiya15:36
rgreeningbeen trying to make a 10.10 netbook image using usb-creator-kde from Lucid. No success. its version
rgreeningev: any ideas?15:37
rgreeningi get init fail15:37
rgreeningcannot find init15:37
rgreeningdrop to initramfs15:37
evrgreening: have you checked that the md5sum of the ISO matches what's on the website?15:38
rgreeningive tried a couple of isos.. daily and beta15:38
evwhat does cat /casper.log say?15:38
rgreening1 sec15:38
rgreeningev: lot's of can't open files or directory15:40
evat boot, can you break into the menu (press a key when you see the keyboard icon) and use the "check CD for defects" option15:41
rgreeningev: it start the splash screen for the test then reboot15:43
rgreeningev: md5sum matches15:45
evrgreening: if you mount the usb disk on a booted system, you can do a md5sum -c /path/to/usb/md5sum.txt15:46
rgreeningsure.. 1 sec15:47
rgreeningev: everything comes back ok15:49
evrgreening: can you pastebin the /casper.log off of there after a failed boot attempt15:50
evjust mount -o remount,rw the usb disk and copy it onto there15:50
rgreeningok. give me a few min. ty15:51
evsure thing15:51
rgreeningev: what's the full cmd?16:06
rgreeningI can't seem to get it to mount16:06
rgreening/deb/sdb1 is the disk16:06
rgreeningev: can't seem to mount /dev/sdb116:09
evrgreening: your disk is likely to be mounted at that point, probably on /cdrom16:10
evmount -o remount,rw /cdrom16:10
rgreeningI made this without persistence. does that matter? Also, nothing shows up for /cdrom in mounts16:11
evrgreening: if it's not already mounted, then mount it in /mnt or whatever.  Are you sure it's sdb1?  Do remember that the disk order can change based on the order your bios gives them, it could be sda1.16:14
rgreening1 sec.. i can verify16:15
rgreeningI have /dev/sda (1,2,5,6 which is the existing system) and /dev/sdb (1) the usb stick I assume.16:16
rgreeninglet me try something else.16:16
rgreeninghmm... ev, acer 1 does something funky16:17
rgreeningIf I have the storage disk inserted, the usb stick fails. if I remove the storage disk (SDHC 16 GB card) then it works16:18
rgreeningI guess the detection picks up the storage disk and fails and never redetects the USB stick16:19
rgreeningnot sure how to address/fix that16:19
evdoubtful.  It sounds like a bug in casper or usb-creator.  Can you please get /casper.log off there from a failing boot attempt and pastebin it?16:20
rgreeningev: when it's failing I can't mount anything.16:20
evwhat error do you get?16:21
rgreeningwrong command I believe. let me try again16:21
evwhat are you typing?16:21
rgreeningmount -o remount,rw /cdrom or with /dev/sdb1/cdrom16:23
rgreeningfirst says not in proc mounts and second invalid argument16:23
rgreening/dev/sdb1 /cdrom16:23
evcan you type out the exact command you're entering.  Not with "or with" in the middle of it.16:24
rgreeningmount -o remount,rw /dev/sdb1 /cdrom16:25
rgreeningmount -o remount,rw /cdrom16:25
evthe first one is bogus16:25
evwhat happens when you type the second one?16:26
rgreeningcant find crdom in /proc/mounts16:26
evrgreening: so mount /dev/sdb1 /cdrom16:26
rgreeninginput/output error16:27
evrgreening: what's the output of mount without arguments16:27
rgreeningrootfs on /, none on /sys /proc /dev /dev/pts and fusectl on /sys/fs/fuse/connections16:28
rgreeningnothing else16:28
evcjwatson: why do you leave the effective UID and GID as the regular user in regain_privileges_save?  This breaks the now following setgroups([])16:29
cjwatsonhm, that looks like a bug16:30
evso 0, 0, 0?16:30
cjwatsonshould be symmetrical with drop_privileges_save16:30
cjwatsonI think so16:30
evI wasn't sure if I was missing some ancient unix rune16:31
evokay, though I am quite curious to know what case you were considering then that we're forgetting now ;)16:31
cjwatsonit was probably just a brainfart16:33
cjwatsonthe only reason I introduced the _save variants was that KDE got pissed off if it saw real uid 016:33
rgreeningev: I was able to mount the SDHC card, just not the USB stick. So I copied the log to the SDHC card and pasted here: http://pastebin.com/Khr0yH9G16:35
evcjwatson: fair enough16:36
evrgreening: what on earth...16:37
evthat looks like the mount point went away16:38
evwhich I guess would be consistent with your IO error16:38
rgreeningev: which I believe is the point at which the system is seeing the SDHC card instead of the USB stick16:39
rgreeningIf I boot without it inserted, it works fine16:40
rgreeningwhich is strange.16:40
evwell, it could've been a bug in casper/usb-creator16:40
evbut that looks unlikely here16:40
CIA-28ubiquity: evand * r4400 trunk/ (debian/changelog ubiquity/misc.py):16:44
CIA-28ubiquity: Set the effective UID in regain_privileges_save so we don't try to16:44
CIA-28ubiquity: setgroups([]) as a regular user (LP: #646827).16:44
rgreeningev: with btrfs work as /boot partition or should I stick with ext type fs16:45
evrgreening: I don't believe the btrfs support in grub landed in time.  I could be wrong though.16:46
* holstein just tried btrfs16:46
holsteingave it a couple days16:46
holsteinactually, it took a day and a half to install ;)16:46
* holstein tried it on an EEEpc16:47
* holstein back to ext2 with no swap :)16:47
cjwatsonev,rgreening: it didn't - blocked on legal16:49
cjwatsonyou can use it as / as long as your /boot is something grub understands16:49
rgreeningcjwatson: ok. thanks. I have / as btrfs and will use ext for /boot16:50
rgreeningcjwatson: what was the 'legal' that blocked it?16:51
rgreeningdiff licences?16:51
cjwatsongrub is GPLv3, the btrfs patch uses stuff from GPLv2-only code, needs Oracle's + Intel's approval to release under terms of GPLv316:52
cjwatsonwe have informal acks from the developers but they've promised something formal and we haven't got it yet16:52
rgreeningah. that makes sense16:52
* holstein made a small ext3 /boot part - 200mb?16:52
holsteinsomewhere in there16:52
rgreeningholstein: mine is 98mb16:52
rgreeningbut I watch and trim the /boot images16:53
cjwatsongenerally seems a bit early for btrfs.  we were enabling it in order to make it easier to help out with development, rather than because we expected real users to use it just yet16:53
rgreeningI'm installing on a test system, so quite ready to roll the dice16:54
holsteinrgreening: did you say solid state drive?16:55
rgreeningholstein: it's an 8GB solid state16:55
holsteinim interested to see how it works for you16:55
rgreeningill know soon enough16:55
holsteinif it goes like my install did, you'll know when the installer starts reporting % complete ;)16:57
rgreeningev: the installer window has a scrollbar. shouldn't it have maximized instead to show the demo during install?16:57
rgreeningthe install window is centered but only takes half the screen width16:57
holsteini remember reading %60, i went to sleep, got up and read %7016:57
rgreeningas a result, the text is cut off16:57
rgreeningunless I scroll over. very unpleasing :)16:58
holsteinrgreening: is btrfs in the live installer now too?16:58
rgreeningholstein: I'm installing from the ubuntu netbook beta and it's there16:59
rgreeningI chose just install ubuntu and not the try ubuntu16:59
* holstein used a maverick alternate daily16:59
cjwatson80GB solid state.  fast *and* you can develop usefully on it. :)17:00
* rgreening likey17:03
rgreeningholstein: install completed. rebooting.17:07
holsteinrgreening: interesting17:07
rgreeningholstein: seems to bott and run fine.17:09
holsteinrgreening: good to know17:10
holsteini bet its something to do with my particual SSD HD then17:10
rgreeningholstein: it's running SLOOOOOOOOW17:27
rgreeningomg.. not sure if its mutter or btrfs or something else17:27
rgreeningi think mutter17:28
holsteinrgreening: :/17:28
* rgreening gonna try the kubuntu one next. compare responsiveness17:28
holsteinseemed like i was experiencing crazy slow write speeds17:29
holsteinresponsiveness seemed fine mostly17:29
rgreeningmutter/X is utterly slow and unresonsive17:30
holsteinyeah, that could be non-btrfs related17:31
ehbellohello guys :)18:46
ehbelloI've noticed about the existence of Burg and I would like to add support for it in the ubiquity installer...18:46
ehbellosomebody can tell me some recommendation to start?18:47
ehbelloI assumesuppose that if BURG is a superset of GRUB, replacing all ocurrences of ubiquity source code I will get my target, but this will remove the support for GRUB :P18:51
ehbelloerh... ubi-summary crashed...20:01
ehbellowhy are bug reports of ubiquity privated?20:02
ehbelloI can't see bug #578130 :-/20:02
ubot2ehbello: Bug 578130 on http://launchpad.net/bugs/578130 is private20:02
superm1ogra, are you seeing debconf communicate crashes after oem-config in your prebuilt images?23:24
superm1seems to happen during the removal of oem-config for me, with both the aptdaemon or the removal method23:25
superm1apport refuses to submit them though (assertion errors)23:26

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!