July 2012 - Anaconda-devel - Fedora mailing-lists

Ability to install to loopmounted filesystem - mentor wanted
by Tirtha Chatterjee 02 Dec '12

02 Dec '12

Hello. I am trying to implement something like Wubi on Fedora. For this, I first need to make it possible to install to loopmounted filesystems through anaconda. I don't think a lot of modification to the existing source code is required. I tried going through the source code, but it seemed quite big to manage without some guidance. I would be extremely happy if someone guided me on this. Hoping for an affirmative reply :-) -- Regards Tirtha Chatterjee KDE developer http://wyuka.co.cc/

2 3

Bill's patch for software selection + comps metadata
by Máirín Duffy 17 Aug '12

17 Aug '12

For those of us not on anaconda-patches list (like myself and the other designers I know are monitoring this list :) ): https://lists.fedorahosted.org/pipermail/anaconda-patches/2012-July/000165.… Here is the screencast Bill Nottingham made of the UI changes in the patch: http://notting.fedorapeople.org/installer.webm Some things Chris, Bill, and I discussed about the patch: - Overall a great idea. - 'Installation Type' is maybe an odd term to use. We talked about using the term 'Environment' instead - that term is inclusive of all the desktop options and (no-desktop) minimal install that are in that selection box. - We use the term 'install class' a lot, so it might be better to pick a different term. Bill has used install class here to mean a desktop + the comps groups that are compatible with it = a base package set. Some alternatives Bill proposed (with some thesaurus help): selection collection combination compilation assembly mix My suggestion is to use "environment collection" (even though it's long) because I think collection is the most appropriate term (a group of sets, think Flickr 'collections' vs 'sets') but it's also important to use the term 'environment' in it so that someone looking at the comps knows that it relates to the 'environment' widget in the UI. Thoughts? ~m

4 7

[PATCH] libudev now has a version of .1
by Mark Hamzy 15 Aug '12

15 Aug '12

libudev bumped the version number up to /usr/lib64/libudev.so.1 Also, change the error message from None. --- pyanaconda/pyudev.py | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/pyanaconda/pyudev.py b/pyanaconda/pyudev.py index 4b5d586..705b93d 100644 --- a/pyanaconda/pyudev.py +++ b/pyanaconda/pyudev.py @@ -31,10 +31,12 @@ def find_library(name, somajor=0): return None # find the udev library -libudev = find_library(name="udev", somajor=0) +name = "udev" +somajor = 1 +libudev = find_library(name=name, somajor=somajor) if not libudev or not os.path.exists(libudev): - raise ImportError, "No library named %s" % libudev + raise ImportError, "No library named %s.%d" % (name, somajor) # load the udev library libudev = CDLL(libudev) -- 1.7.10.1

4 3

[newui] Remove scdate usage and have timezone only in ksdata
by Vratislav Podzimek 07 Aug '12

07 Aug '12

PATCH 1/2 removes the last piece of live code that uses the system-config-date (note that there still is the old GUI code using the scdate's timezone map). However, I didn't find scdate as a dependency in anaconda.spec.in nor in lorax templates for the installation runtime environment. How it gets to the install image? PATCH 2/2 does basically two things: 1) It removes the Timezone class used in the anaconda object for timezone data replacing it with the code using ksdata as the only sctructure storing data and execute method of the Timezone pykicstart command. 2) It moves some pieces of code from the localization module to the separate timezone module adding a function for writing timezone configuration. This function is based on the old one from the Timezone class. I have just replaced copying timezone file to the /etc/localtime with creating a symlink because this way it is easy to find out which timezone the system is using. -- Vratislav Podzimek

4 10

[PATCH] Some accessibility improvements
by Joanmarie Diggs 06 Aug '12

06 Aug '12

Hey all. As some of you know, I have been threatening to do this work for a while now, but other things keep bubbling to the top of my priority list. :-/ I'm really sorry about that. Rather than wait until I have everything done, I'm attaching some changes which greatly improve the accessibility of Anaconda for users who are blind and are, insofar as the code goes, pretty trivial. What the patches do is: 1. Set accessible relations between labels and the things being labeled. As a result, instead of Orca saying things like "text" and leaving the user to guess what text he/she should type, Orca now says "foo: text" and the user knows to type a value for foo. This was also done for labels associated with a group of widgets so that the user has some context when navigating. Note that there are a couple of ways you can tackle this: Add accessibility-specific code to set the relations or just set the mnemonic widget property and let Gtk do it for you automatically. I wound up doing the latter based on the suggestion of Benjamin Otte (Gtk+ developer). 2. Change initial focus for each* installer step to the likely widget of interest. As a result, instead of Orca merely saying "Next" and leaving the user to wonder what step he/she is on and what is required in order to proceed, Orca now automatically presents the widget and associated context for each new screen/step that appears ("Step foobar, you must enter a foo and a bar, foo text"). * Well, most steps. More work is needed here. 3. Because the change in step 2 in some cases broke the "Just press Return for the 'Next' button" functionality, make the 'Next' button the default. That brings back the "Just press Return" functionality so the end user experience for everyone else should still be the same as it currently is. If you have any questions, please let me know. Thanks, take care, and again my apologies for the delay. --joanie (Orca project lead)

3 3

Key installer functionality for F18: rescue mode, text/serial console mode
by Adam Williamson 02 Aug '12

02 Aug '12

Hey, folks. Since there's been some discussion on test@ about possibly-missing features with the f18 newUI transition, I wanted to make sure we're all on the same page about what needs to be in for f18. the two features QA was worried about missing are text mode installation, and the rescue mode that's present on the DVD/netinst images. on text install: apparently it's now the plan that a basic text mode will be included in f18 and may even make Alpha. As long as that's a commitment then that should work out fine and is no longer a worry. on rescue mode: we would like to see a roughly functional equivalent replacement for rescue mode present in f18, with considerable flexibility as to the exact form. as far as I can tell, these are the major features of rescue mode: 1) bring up network 2) mount installed system partitions 3) provide rescue shell with binaries from installed system in $PATH it also has a 'diagnostics' thing, but that seems somewhat obscure - it gives you a screen full of cryptic output when you run it, but that output almost all appears to come from some sort of mdraid analysis thingy. It doesn't seem to do much beyond that. Anyone have more details on the 'diagnostics' feature and how it's supposed to work? Anyway, diagnostics aside, it shouldn't be too much effort to implement something that does 1, 2 and 3. One idea has been to make it as a module for dracut. wwoods suggested "personally what I'd want to do is use lorax and make a rescue runtime, or just add stuff to a rescue target in the normal anaconda runtime". I'm not making any suggestion for what method would be best - anything works so long as there's something =) If anaconda team can commit to a rescue mode replacement for f18 final at least, that'd be great. thanks. Also, if anyone's aware of any other major functionality that might go away with newUI - besides what's already been discussed or called out on the feature page - please let us know. thanks! -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Twitter: AdamW_Fedora | identi.ca: adamwfedora http://www.happyassassin.net

9 23

Bug in newui and python-meh problem
by Vratislav Podzimek 01 Aug '12

01 Aug '12

When trying to make python-meh exception handling work, I've hit a traceback [1]. It happens every time, but since it is in separate thread, without python-meh everything else goes on. I've used the updates.img to use up to date code. Another problem is, that if unhandled exception appears in a separate thread, python-meh displays the mainExceptionWindow, but then immediately whole Gtk crashes with many error messages on tty1. I guess the problem is that python-meh runs another Gtk main loop from a different thread. Any ideas how to resolve that? Appart from that, I added some patches that make python-meh work (posted on our patches list). [1] http://fpaste.org/Fbiu/ -- Vratislav Podzimek Anaconda Rider | Red Hat, Inc. | Brno - Czech Republic

4 4

Text based UI - framework core, example hub&spokes and input related questions
by Martin Sivak 31 Jul '12

31 Jul '12

Hi, Lets move this to devel list so others can comment too. > That's fair. I really don't know how many of those spokes we can > actually do in text mode given the constraints though. So the current info I got from David is that we are required to have full featured text mode. We can do spokes which consist of more linear screens though to make up for the lack of screen space. > > > > Your examples used the Hub & Spoke model and you need a way of > > returning back to the Hub when the Spoke is done. We also need > > linear > > progression (to move between hubs or inside Spoke screens) and two > > "dialogs" (showError and showYesNoQuestion as required by > > UserInterface base class). > > This is true. My example was just about showing the UI, not about > how > to actually code it. As far as navigation is concerned, we certainly > want to re-use as much of the existing ui code that's already in > ui/gui/. I agree and we should be able to separate the UI agnostic parts to some kind of common/ abstract classes and then inherit from those for GUI and TUI cases. > I agree that columns are a bad idea. Particularly when you start > throwing translated strings into those. Well not really, this might still work, try the example code I did yesterday (I am attaching it again, since you probably missed it on devel list). I do not plan on adding much more widgets now, but what is there can actually do columns pretty easily and it will work even on dot matrix printer with translations in place :) > I don't necessarily have something different in mind. I am > struggling > with how to make use of this code with a real spoke. Right now I'm > looking at what it takes to re-use much of the code existing in > ui/gui > to setup a single text hub and single text spoke and somehow get it > displayed. Your code may in fact become handy there, something in > place > of Gtk.main(). That is the idea behind App and UIScreen classes. I am reading input using raw_input so no single keypresses at the moment, but that might be easily changed. There is one more issue coming up, how do handle textual/numeric input? Can we use readline on s390? Or maybe use readline only where it is supported? Does anybody know how does that work on those dumb terminals? Martin

2 14

Feature request: ipmi
by Jeffrey_L_Mendoza＠Dell.com 31 Jul '12

31 Jul '12

Hi List, I have a feature request for https://fedoraproject.org/wiki/Anaconda/Features : ipmi functionality. An immediate feature would be the ability to include freeipmi-bmc-watchdog. This would allow administrators to use the service processor/bmc to detect/report a stuck install. Future use would be to have anaconda run ipmi commands to log install status: start, step completed, success, failure. Therefore having the service processor aware of the install status. I assume this is accomplished by including the ipmi driver and ipmitool/freeipmi packages in the stage 2 image. I am in the process of familiarizing myself with the wiki: http://fedoraproject.org/wiki/Anaconda/SourceOverview http://fedoraproject.org/wiki/Anaconda/Stage2DevelopmentGuide Thanks, Jeff Mendoza Software Engineer Dell | Linux OS Engineering

3 2

Use systemd-localed?
by Vratislav Podzimek 31 Jul '12

31 Jul '12

As it was suggested in the discussion here on the devel list, we could use the systemd-localed DBus service [1] to set keyboard layouts. I've looked at it and tried some tests and here are the results: 1.) The main problem is, that this service allows to set only one keyboard layout and is therefore not usable for us for setting multiple X11 keyboard layouts. 2.) While it has a method for setting the virtual console keyboard layout too, it provides us the same functionality as loadkeys command. I see no reason for replacing loadkeys calling with DBus method invocations. Once rhbz#837292 [2] gets resolved it would allow us to use more keyboard layouts in the console, but for now I don't see any reason to use it. Comments are, of course, welcome. [1] http://www.freedesktop.org/wiki/Software/systemd/localed [2] https://bugzilla.redhat.com/show_bug.cgi?id=837292 -- Vratislav Podzimek Anaconda Rider | Red Hat, Inc. | Brno - Czech Republic

2 4

newui networking todo
by Radek Vykydal 27 Jul '12

27 Jul '12

I put newui networking todo, notes, questions here: https://fedoraproject.org/wiki/Anaconda/Network#Newui_networking_TODO Radek

1 0

[rhel5-branch] Syslinux changed paths from /usr/lib to /usr/share (#843290)
by Martin Gracik 27 Jul '12

27 Jul '12

--- scripts/mk-images | 2 +- scripts/mk-images.x86 | 6 +++--- scripts/upd-instroot | 2 +- 3 files changed, 5 insertions(+), 5 deletions(-) diff --git a/scripts/mk-images b/scripts/mk-images index 61dc004..4823ad0 100755 --- a/scripts/mk-images +++ b/scripts/mk-images @@ -886,7 +886,7 @@ done doPostImages # do some clean up before making the cramfs of the second stage -rm -rf $INSTIMGTEMPLATE/usr/lib/anaconda-runtime $INSTIMGTEMPLATE/usr/lib/syslinux $IMGPATH/usr/lib/syslinux +rm -rf $INSTIMGTEMPLATE/usr/lib/anaconda-runtime $INSTIMGTEMPLATE/usr/share/syslinux $IMGPATH/usr/share/syslinux if [ -n "$foundakernel" ]; then # make the second stage diff --git a/scripts/mk-images.x86 b/scripts/mk-images.x86 index 38556f6..d56d782 100644 --- a/scripts/mk-images.x86 +++ b/scripts/mk-images.x86 @@ -1,4 +1,4 @@ -SYSLINUX=$IMGPATH/usr/lib/syslinux/syslinux-nomtools +SYSLINUX=$IMGPATH/usr/share/syslinux/syslinux-nomtools if [ ! -f $SYSLINUX ]; then echo "Warning: nomtools syslinux doesn't exist" @@ -67,12 +67,12 @@ makeBootImages() { --modules "$INITRDMODS" [ $? = 0 ] || exit 1 - if [ -f $IMGPATH/usr/lib/syslinux/isolinux.bin ]; then + if [ -f $IMGPATH/usr/share/syslinux/isolinux.bin ]; then echo "Building isolinux directory" MBD_BOOTTREE=$TOPDESTPATH/isolinux MBD_FSIMAGE=$TOPDESTPATH/images/pxeboot/initrd.img mkdir $MBD_BOOTTREE - cp $IMGPATH/usr/lib/syslinux/isolinux.bin $MBD_BOOTTREE/isolinux.bin + cp $IMGPATH/usr/share/syslinux/isolinux.bin $MBD_BOOTTREE/isolinux.bin prepareBootTree diff --git a/scripts/upd-instroot b/scripts/upd-instroot index d83d375..b18675d 100755 --- a/scripts/upd-instroot +++ b/scripts/upd-instroot @@ -775,7 +775,7 @@ usr/$LIBDIR/python?.?/site-packages/cairo/* usr/$LIBDIR/python?.?/site-packages/gtk*/gtk/* usr/$LIBDIR/python?.?/site-packages/gtk*/* usr/$LIBDIR/rpm/rpmpopt -usr/lib/syslinux/* +usr/share/syslinux/* usr/lib/yaboot/* usr/sbin/chroot usr/sbin/ddcprobe -- 1.7.9.5

2 1

First mock up of new text UI
by Jesse Keating 25 Jul '12

25 Jul '12

I spent today thinking through text installs in newui. While I believe that text installs should have minimalistic configuration options, I also believe that the little it has should follow the same sort of idea as the hub/spoke design of newui. Another requirement of text mode is that the same UI work across all text interfaces (real tty, serial, ssh, s390 x3270 shell, etc...), targeting the lowest common platform. For systems like s390x that means no curses, no fancy screen widgets, just dumb text printed on a somewhat narrow screen with no scrolling capability. What I've done is a hacky mock up of how a text install would progress. Some of these config options might go away (like lang/keymap since we have good command line argument options for these). The idea is the same as newui. A hub from which configuration can be done in any order, but also with some items that are required to be completed. Of note, I brought back the root password selection because with text mode you only get @minimal, and thus no firstboot. We don't want people to have to root their machines just to get a password set. This is pretty much the cmdline UI slightly re-designed. You'll find a series of screen shots at http://jkeating.fedorapeople.org/textui/ Walk through them as numbered (yes I'm missing a 2-...). I have no idea yet if we can handle many choice screens like lang and keyboard. Timezone data can be broken down much how tzselect does it. Feedback would be appreciated! -- Jesse Keating Fedora -- Freedom² is a feature!

6 13

mdraid DDF container support
by Jared_Dominguez＠DELL.com 25 Jul '12

25 Jul '12

Greetings, I've been working on a feature to add DDF container support using mdraid to Anaconda. It should be building on the work done to support IMSM containers using mdraid instead of dmraid. I've been hoping to have a patch ready already, but I'm running into some problems related to differences in how metadata for the two container formats is exposed. This is also the first time I've worked on Anaconda so am still familiarizing myself with the codebase. For IMSM, Anaconda has a udev rule that gets a value for MD_DEVICES: ENV{ID_FS_TYPE}=="isw_raid_member", IMPORT{program}="$env{ANACBIN}/mdadm --examine --export $tempnode" Alas, for DDF, the same rule doesn't get a value for MD_DEVICES. One can get this information for DDF by first doing "mdadm -A -e ddf --scan". Subsequently, "mdadm --detail --export /dev/md127" will give MD_DEVICES. However, it seems to me to be a poor practice to assemble any arrays in 70-anaconda.rules. Besides, while testing, Anaconda seems not to like this in there anyway: ENV{ID_FS_TYPE}=="ddf_raid_member", RUN+="$env{ANACBIN}/mdadm -A -e ddf --scan" KERNEL=="md[0-9]*|md_d[0-9]*|md/*", IMPORT{program}="$env{ANACBIN}/mdadm --detail --export $tempnode" It interferes with Anaconda accessing the array with mdadm, causing Anaconda to hang indefinitely. Might anyone with more Anaconda development experience have suggestions for where to try collecting this one piece of information? --Jared -- Jared Dominguez Linux Engineering Dell | Enterprise Product Group

2 2

glade files
by Chris Lumens 25 Jul '12

25 Jul '12

I made a mistake when adding these glade files. I should have given them the default .glade extension instead of .ui like I thought was the standard. xgettext requires that they have a .glade extension to be picked up for translation. So, I'm going to push a patch that moves everything. I hate having to do this because it screws up git history unless you use the --follow option. On the other hand, who ever looks at the git history of a UI file? - Chris

1 0

anaconda disk image partitioning question
by David A. Marlin 25 Jul '12

25 Jul '12

I am using livemedia-creator on F17 to create a disk image. I manually select the partitioning scheme in the kickstart: clearpart --all part /boot --size 255 --fstype ext3 --label=boot part swap --size 511 part / --size 3327 --fstype ext4 --label=rootfs Previously I created several disk images and they all had the partitions in the following order: 1 - /boot 2 - swap 3 - / (root) I created another disk image today, but the partitions are now in the following order: 1 - /boot 2 - / (root) 3 - swap The reason this is significant is that after installing the disk image onto physical media I want to be able to resize the root partition ('/'), and that is only practical if it is the last partition on the image. So my question is, what determines the order of the partitions on a disk image, and is there a way to select, or guarantee a specific order? Thank you, d.marlin

2 2

Release criteria vs. Anaconda
by Petr Schindler 23 Jul '12

23 Jul '12

Hi, I'd like to ask you. Which interfaces you want to stop support? I've read that there won't be text interface. Are there any other changes which we should reflect in release criteria [1][2][3]? [1] https://fedoraproject.org/wiki/Fedora_17_Alpha_Release_Criteria [2] https://fedoraproject.org/wiki/Fedora_17_Beta_Release_Criteria [3] https://fedoraproject.org/wiki/Fedora_17_Final_Release_Criteria

2 1

[PATCH anaconda/master] Add support to determine the ARM processor variety and select the correct kernel to install.
by David A. Marlin 20 Jul '12

20 Jul '12

Attached is my latest version of the ARM processor variety patch. It incorporates suggestions made previously, and was made against anaconda/master (but applied to 17.29 for testing). I simplified the hash/lookup by consolidating some machine types, but I understand this may still be unacceptable. I am currently tracking master and carrying these patches locally. I keep the rpms in a local repo for testing. We are working on a method of exposing the platform in userspace, but that is not in place yet. Would it be possible to accept this as an interim solution with the expressed intent that the hash would be removed as soon as an alternative solution is available? If not, would a version of this patch with the hash/lookup stubbed out be acceptable? I would like to get as much of this upstream as possible, to minimize the patch(es) I need to carry locally and simplify keeping up with master. Any suggestions are appreciated. Note: We successfully completed a test kickstart install on an ARM system today using the modified anaconda-17.29. d.marlin

6 12

[PATCH] Re-remove fedora-import-state
by Will Woods 18 Jul '12

18 Jul '12

Commit c8c14c9 deleted these two files, but they got resurrected in some merge somehow. Delete them again, since we don't need them. --- data/systemd/fedora-import-state | 14 -------------- data/systemd/fedora-import-state.service | 13 ------------- 2 files changed, 27 deletions(-) delete mode 100644 data/systemd/fedora-import-state delete mode 100644 data/systemd/fedora-import-state.service diff --git a/data/systemd/fedora-import-state b/data/systemd/fedora-import-state deleted file mode 100644 index a853ab9..0000000 --- a/data/systemd/fedora-import-state +++ /dev/null @@ -1,14 +0,0 @@ -#!/bin/bash -# fedora-import-state: import state files from initramfs (e.g. network config) - -# exit early if root isn't writeable -[ -w / ] || exit 0 - -# copy state into root -cd /run/initramfs/state -cp -a -t / . - -# run restorecon on the copied files -if [ -e /sys/fs/selinux/enforce ]; then - find . | ( cd /; restorecon -i -f -; ) || : -fi diff --git a/data/systemd/fedora-import-state.service b/data/systemd/fedora-import-state.service deleted file mode 100644 index 5294fe1..0000000 --- a/data/systemd/fedora-import-state.service +++ /dev/null @@ -1,13 +0,0 @@ -[Unit] -Description=Import network configuration from initramfs -DefaultDependencies=no -ConditionDirectoryNotEmpty=/run/initramfs/state -Conflicts=shutdown.target -Before=shutdown.target emergency.service emergency.target systemd-tmpfiles-setup.service -After=remount-rootfs.service - -[Service] -ExecStart=/lib/systemd/fedora-import-state -Type=oneshot -TimeoutSec=0 -RemainAfterExit=yes -- 1.7.10.4

1 1

[PATCH] Re-remove fedora-import-state
by Will Woods 18 Jul '12

18 Jul '12

Commit c8c14c9 deleted these two files, but they got resurrected in some merge somehow. Delete them again, since we don't need them. --- data/systemd/fedora-import-state | 14 -------------- data/systemd/fedora-import-state.service | 13 ------------- 2 files changed, 27 deletions(-) delete mode 100644 data/systemd/fedora-import-state delete mode 100644 data/systemd/fedora-import-state.service diff --git a/data/systemd/fedora-import-state b/data/systemd/fedora-import-state deleted file mode 100644 index a853ab9..0000000 --- a/data/systemd/fedora-import-state +++ /dev/null @@ -1,14 +0,0 @@ -#!/bin/bash -# fedora-import-state: import state files from initramfs (e.g. network config) - -# exit early if root isn't writeable -[ -w / ] || exit 0 - -# copy state into root -cd /run/initramfs/state -cp -a -t / . - -# run restorecon on the copied files -if [ -e /sys/fs/selinux/enforce ]; then - find . | ( cd /; restorecon -i -f -; ) || : -fi diff --git a/data/systemd/fedora-import-state.service b/data/systemd/fedora-import-state.service deleted file mode 100644 index 5294fe1..0000000 --- a/data/systemd/fedora-import-state.service +++ /dev/null @@ -1,13 +0,0 @@ -[Unit] -Description=Import network configuration from initramfs -DefaultDependencies=no -ConditionDirectoryNotEmpty=/run/initramfs/state -Conflicts=shutdown.target -Before=shutdown.target emergency.service emergency.target systemd-tmpfiles-setup.service -After=remount-rootfs.service - -[Service] -ExecStart=/lib/systemd/fedora-import-state -Type=oneshot -TimeoutSec=0 -RemainAfterExit=yes -- 1.7.10.4

1 0

noloader: create default ifcfg files already in dracut
by Radek Vykydal 18 Jul '12

18 Jul '12

Hi, this proposition is mainly for Will. Currently we create missing default ifcfg files (i.e. file of devices not activated in dracut or configured by kickstart) in anaconda: commit e1c287f31dcec4c661272f852e14f5d6925f92d2 Author: Will Woods <wwoods(a)redhat.com> Date: Wed Mar 21 20:02:46 2012 -0400 Create default ifcfg-* for each interface (#804504, #804716) Loader used to create a default ifcfg-* file for each unused network interface, containing DEVICE, HWADDR, UUID, BOOTPROTO=dhcp, etc. Network.update() populated self.netdevices by reading those files, *skipping any interface that lacked that file*, which caused the vnc traceback in bug #804504. And without BOOTPROTO=dhcp, "ifup $dev" didn't work post-install, which caused #804716. (Personally I think "ifup" should fall back to the same defaults as NetworkManager rather than making us write the defaults to a file, but we'll save that for later.) This has a problem that NetworkManager activates default connections for devices lacking ifcfg file which differs from behaviour in rhel6 and I think we don't want to start to do this. Perhaps we could run anaconda pre-pivot hook (post pre-pivot/85write-ifcfg.sh) that would take care of missing ifcfg files. What do you think? Should I try to come up with an anaconda dracut patch? Also I'd like to bring back functionality of --activate option, but activation of additional devices could be done in anaconda kickstart processing, even though doing that in dracut might be as simple as writing ONBOOT=yes by dracut/ks to ifcfg file. Radek

2 3

UI status braindump, part I
by Máirín Duffy 16 Jul '12

16 Jul '12

Hi! CC'ed are Ryan Lerch (new Fedora UX designer who just relocated to Westford from Brisbane) and Mackenzie Colwell (whom you've likely already met as pseudomonas in IRC, she's my UX summer intern.) We met with Chris this morning (and will meet further later this afternoon) to hash out the current status of the UI design. Here's a rundown of what we covered this morning; I'll follow up with a part II for what we cover this afternoon. #1) KICKSTART DETECTION ======================== So we've had this idea that if you have a USB key with a .ks file on it when the Anaconda UI is started, we would auto-detect it and offer to pre-fill in all the UI fields (and add post scripts and whatever else into the fray) for the install. Mackenzie worked up some mockups (yet to be posted) showing how this might look and feel. Some notes: - If we make this feature infinitely flexible now, it's going to be very hard to whittle it down later because people may come to rely on different aspects of it. So we discussed limiting the scope of the feature initially and expanding eventually based on user feedback. - We talked about only auto-detecting ks files that are: - in the root directory of a USB key - in the root directory of the install media - We made the assumption that a user is highly unlikely to have more than 0-10 ks files auto-detected, and the UI mockups are designed accordingly. - We talked about a feature (that could be delayed release-by-release depending on time / resources available to implement) to allow users to pull in only certain named sections of a KS file. So, we auto-detect the KS file and that it has a %packages and a %post. You could dive in an select to only pull in the %packages section with checkbox, and uncheck the %post to not have it apply to your install at all. - If the stanza-by-stanza pick and choose for KS feature isn't implemented, we default to all or nothing when we auto-detect ks files: you either have all parts of the KS applied to your install or none. - For picking and choosing sections of a ks file, we determined that we should have all sections included by default, and allow the user to 'opt-out' of particular sections. (rather than vice-versa where they would start with none and have to opt-in one-by-one.) - We talked about how opting in to the %package section of the ks file might be a good part of the story for folks concerned that we have dropped individual package selection from the UI. - A kickstart file with just one section will pass validation. Old style kickstarts default to asking for any information not provided in the KS. In the new anaconda UI, we will pre-fill in the defaults for any fields the KS did not provide, and if any section is not provided we will put the user on the 1st hub rather than the 2nd hub. (Not sure if we should push users to the 2nd hub if every required KS option is filled out in the KS file sucked in.) - For picking and choosing KS files, we need a way to view each section so you know what you're signing up for (or know what you're avoiding by unchecking it.) #2) CHOOSE DISKS BEFORE HUB #1 ============================== http://linuxgrrl.com/fedora-ux/Projects/Anaconda/Flows/full-flow_26sep2011.… I know we came up with the "Choose Disks" flow in there to limit the number of disks selectable later on in the UI to 10 or less because of a discussion we had in Westford some months back. I don't remember all the particulars (eek) so I wonder if we still need it? If so we need to mock it up. #3) LOCAL MEDIA PREUPGRADE ========================== We want to choose preupgrade as the one true / supported upgrade path. However! This is problematic for folks with crappy, expensive, or no internet connections. Can we investigate whether it would be possible to offer up a 'preupgrade from DVD media' option? What if the DVD media was live, possible? #4) KEYBOARD LAYOUT IN LANG SELECTION SCREEN #1 =============================================== So, it's been pointed out that while we allow users to select the language the Anaconda UI is displayed in on the very first screen, we don't allow them to select a keyboard layout. This is problematic because we have a network dialog after that, and if they cannot type their AP password or input their network information using the default layout for their chosen language, they can't benefit from any of the geo-ip and network features (like pre-downloading repo metadata) the new UI offers. Ryan has been looking at integrating keyboard layout selection into this screen. As we discussed this a number of issues came up: - We likely won't have geo-ip data for this first screen so it would be really difficult to filter based on geo relevancy, but we can do this on the language screen off of hub #1. - A user may select one language for the display and another language for input. An example of this is an American consultant stationed on-site at a Japanese business, where the AP password is in Japanese but the American (speaking English natively) would prefer to read the Anaconda UI in English.) - Because of the above point, we should display layouts relevant to the selected language first on screen #1, but also allow the option to add any other layout outside of that language. (with type-ahead) - Will ibus work in Anaconda? If not, CJKVI (Chinese, Japanese, Korean, Vietnamese, Indic, maybe others'?) languages' input won't be possible in Anaconda. (We have had a specific user requests to support Japanese input in anaconda so it is needed / useful.) - For ibus languages, we need to indicate what the key combo is to select between 'sublayouts' (eg. for Japanese, switching between Romaji / Hiragana / Katakana.) #5 UP-FRONT NETWORK AUTO-DETECTION ================================== We had a worrisome and complicated discussion of how to handle auto-detected / auto-connected network support (before hub #1) until Ryan essentially pointed out that NetworkManager has a workable recipe for auto-connecting to a network and we should probably just use whatever scheme NM uses :) Diagram: http://www.flickr.com/photos/mairin/7557284668/in/photostream - If you have a wired connection (that can access the internet not just a local network), you don't see any network config screen before hitting hub #1. - If you have a wireless connection, we pop up a 'simple dialog' (to be mocked up by Ryan) where you pick an access point and input the password (if necessary) - If you have a more complicated connection, you can visit the full blown network dialog (before hub #1) by clicking on a 'Network Settings' button in the 'simple dialog' Some use cases we talked through: - You're at a friends house. They have a MAC filter or password. - You're at the office. All the visible APs are password-protected (maybe using RSA tokens or certs) - You're at a hotel / airport / conference show floor and the network has a captive auth system. You are only connected to a local network then, and not connected to the full internet. - Fake 'Free Public Wifi' connection that has no internet uplink. - You need to connect to a phone's network connection via wired (USB) or wireless tether. - You're at a coffeeshop and the AP is completely open, but you'd rather connect to another AP. If you switch APs, the yum repo metadata download will have to be restarted. Other points - - We could detect whether or not APs were connected to the wider internet or not. TO-DOs for people :) Mackenzie ========= - update the ks auto-detect mockups to allow users to view each stanza of the ks on the same screen where they can check / uncheck stanza-by-stanza - update the ks auto-detect mockups to drop the 'choose sections' button. instead of having a separate 'choose sections' screen, integrate the choose sections widgets into the main dialog using a disclosure widget (example: http://xkahn.zoned.net/blog/wp-content/uploads/2007/02/evolution-compose-te… the 'Show Attachment Bar' in the bottom left of this SS uses a disclosure widget - also called GtkExpander - but in GTK3 they look like little [+] symbols instead of a triangle.) Ryan ==== - (deferred until complete set available) talk to Brisbane docs contents about an audit of the language used in the UI strings to make sure it's internally consistent and consistent with RH style - Update screen #1 mockups to include keyboard layout selection - On keyboard layout screen, for CJKVI languages, provide the key combo for switching between 'sub layouts' (katakana vs hiragana vs romaji in Japanese for example) - Talk to Brisbane contacts about state of ibus and whether or not any alternative input systems are under consideration or if any pending new ibus features are coming up so we're aware - Mock up 'simple dialog' for network connections Mo == - I need to talk to dlehman about the 'choose disks' box we have in the flow chart and whether or not it's still necessary (eg http://linuxgrrl.com/fedora-ux/Projects/Anaconda/Flows/full-flow_26sep2011.… ) Chris & other Anaconda riders ============================= - Search the install media and any USB media attached to the system for ks files, only in the root directory. - Modify kickstart.py as necessary so that it can determine whether or not a given file passed to it actually is a ks file. - If a ks file is loaded in by a user, always take them to hub #1 by default, even if you have sufficient info to start the install. - (low priority) Investigate what it would take to list what ks stanzas / sections are present in a ks file and optionally, section-by-section, turn them on or off. - (dependent on last item above) When pulling in a user-selected ks file, set all ks sections to ON by default. - Investigate preupgrade off of local media feasibility - Investigate ibus in anaconda feasibility - Change code around popping up network dialog before hub #1 to match rules described in section #5 above. Okay more later hopefully :) Any questions etc please reply to list! ~m

5 8

[PATCH] add command line option to override the platform that we are installing for passed onto anaconda to make ARM images where the compose hardware does not match the target hardware
by Dennis Gilmore 16 Jul '12

16 Jul '12

Signed-off-by: Dennis Gilmore <dennis(a)ausil.us> --- src/sbin/livemedia-creator | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/src/sbin/livemedia-creator b/src/sbin/livemedia-creator index 453343a..13d059e 100755 --- a/src/sbin/livemedia-creator +++ b/src/sbin/livemedia-creator @@ -554,6 +554,8 @@ if __name__ == '__main__': help="Keep raw disk image after .iso creation" ) parser.add_argument( "--no-virt", action="store_true", help="Use Anaconda's image install instead of virt-install" ) + parser.add_argument( "--platform", + help="the platform to use when creating images for ARM" ) parser.add_argument( "--proxy", help="proxy URL to use for the install" ) parser.add_argument( "--anaconda-arg", action="append", dest="anaconda_args", @@ -740,6 +742,8 @@ if __name__ == '__main__': anaconda_args += arg.split(" ", 1) if opts.proxy: anaconda_args += [ "--proxy", opts.proxy ] + if opts.platform: + anaconda_args += [ "--targetplatform", opts.platform ] # Use anaconda's image install install_error = anaconda_install( disk_img, disk_size, opts.ks[0], -- 1.7.10.4

1 0

newui python-meh
by Vratislav Podzimek 16 Jul '12

16 Jul '12

Here is a video preview of the new python-meh design implementation: http://vpodzime.fedorapeople.org/python-meh_preview.ogg It follows the debug-off mockup [1]. There also is a debug-on mockup [2], but I think adding the debug option and the code for it just to have a Debug button "one click closer" is not necessary. Any comments or ideas? [1] http://duffy.fedorapeople.org/temp/python-meh/unhandled-debug-off-expanded.… [2] http://duffy.fedorapeople.org/temp/python-meh/unhandled-debug-on-expanded.p… -- Vratislav Podzimek Anaconda Rider | Red Hat, Inc. | Brno - Czech Republic

4 14

Regarding installer development
by shubham 14 Jul '12

14 Jul '12

Hi All, This is my first mail to this ML. I am a newbie in installer development. I want to learn all the activities which are required for development of an installer. Is there any guide /tutorial or link from where I can get in depth knowledge of installer development from some upstream OS . It would be really helpful if some links of tips and tricks are also shared. Thanks in advance. Regards, S.S

1 0

UI status braindump, part II
by Máirín Duffy 13 Jul '12

13 Jul '12

As promised, here's part II of the UI status braindump, starting from where we left off with #6.... 6) Show passwords checkbox for password fields ============================================== Sort of like this screenshot from network manager: http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/6/html/Installat… See the 'show password' checkbox? This would be useful to have on password fields before hub #1 in case authentication to an access point fails - it will allow you to make sure the password isn't wrong because you have the wrong keyboard layout set. Ryan verified that we already have a 'show key' box on the wep key dialog in the mockups now. When we're closer to a complete / integrated set (right now each of us is working on our own copy of the mockups and haven't merged them yet) he'll do a run-through to make sure all of the fields that need this feature are noted in the mockups. 7) Fedora Software Selection ============================ For Fedora, we talked about: - Left column, your option of desktop spins as well as a minimal "no desktop" option equivalent to @core - Right column, functional spins package groups (e.g., electronic lab, security spin, design suite, etc.), normal package groups without language support groups. - adding extra repos - still needs to be worked out, we forgot to talk about this. 8) Partition Shrinking ====================== Eg http://blog.linuxgrrl.com/2012/04/18/rough-thoughts-on-reclaiming-space-fro… This is waiting on me producing a mockup. I talked to dlehman in IRC about this last week. I had a conversation with Garrett Lesage (another open source designer) showing him our mockups so far and he gave the feedback that we should just have one slider per device. The slider might have various 'sticky points' where more drastic actions happen the more you shrink. E.g., say you need 4 gb to do the install. Say you have a pre-existing windows installation on the device with three partitions: - partition A: windows_c with 2 GB free on filesystem - empty space: 2 GB between parts A & B - partition B: windows_d with 0MB free on filesystem - partition C: windows_e with 2 GB free on filesystem You'll have 2 points on the slider maybe: - slide to point 1 and you'll meet the 4gb requirement. You'll resize partition A to free the 2 GB, which is adjacent to the 2 GB empty space between partitions A & B. - slide to point 2 and you'll free up an additional 2 gb by resizing partition C down 2 GB. lvm or btr could be used to make it part of the same logical volume as the chunk of space on the other side of partition b. So there's a preference for resizing partitions that have bigger potential chunks of space to free (so we use partition A first rather than partition C, because even though they are the same size, partition A has a bigger potential being adjacent to a large free space than partition C does.) If you want more space for the filesystem though, moving the slider further does the more drastic action of using a logical volume to chain the two physically disparate chunks together into one for your fedora install. 9) Storage auto-default if Fedora and if only one non-removable disk detected ==================================================================== This is an idea Ryan had when he first reviewed the Anaconda mockups and got started on the project. Right now, no matter what, you can't click past the hub #1 - you have to fill out information in the storage hub. However, a big chunk of Fedora users are setting up single-user Fedora installations on laptops that have a single hard drive. So, if we auto-detect that there is only one non-removable hard drive on a system, we could allow these users to click past the storage screen by automatically selecting the single laptop drive for install. This doesn't mean we wipe that single disk. Instead, we autopart any pre-existing free space on the disk. If that isn't enough space, we automatically shrinkany filesystems as needed. If that isn't enough, we dump 'em in hub #1 with the 'attention' ! icon on the storage part and the user can sort it. :) 9) Icons for specialized disks =============================== All the advanced / network storage disks have the same network storage icon. It might be good to be able to distinguish between raid devices, sans, dasd/zfcp, etc. So we could find and as necessary develop new icons and hook them into the different types of storage when they're displayed. 10) Subscription stuff ====================== This is something that has to be sorted out with the subscriptions team, as well as the folks working on the F18 Initial Experience feature. :( One point that Ryan made was that the GNOME first experience mockups are very user-centric while firstboot is very system-centric and not particular to a specific user much. 11) Bug-reporting / Debug opt-in ================================ Jiri Moskovcak brought this up in a recent email thread with me and Chris - the main ide ahere is that it would be nice to let the user set various debugging options at install time. So they could opt in to ABRT and installing debuginfo packages, for example. I need to talk to Jiri more and get more information about this because I'm still a bit confused about the details. Some things that came up when Chris, Ryan, Mackenzie, and I talked about it today: - opt-in should be someone off hub #1 because it affects whether or not abrt gets installed and particular debuginfo packages get installed - what if it was in firstboot, and abrt / debuginfo was installed in a post-script? or queued up in packagekit? - could they be queued up in the offline updates system? - is there any mechanism to autoclean debuginfo packages when they are stale and useless? - smolt has a new name... this might change the smolt mockups we have - for oem usages of firstboot, there is a reconfig option in the old firstboot, it should carry on to the new one. 12) Random point ================ - How are we handling pre-existing LUKS passwords in the new ui? I gotta go home so no time to do the to-do list summary, maybe tomorrow. :) ~m

6 9

[PATCH lorax/master] Add 'mvebu' to list of recognized ARM platforms.
by d.marlin 12 Jul '12

12 Jul '12

From: "d.marlin" <dmarlin(a)redhat.com> Signed-off-by: David A. Marlin <dmarlin(a)redhat.com> --- share/arm.tmpl | 1 + share/runtime-install.tmpl | 2 +- src/pylorax/treebuilder.py | 2 +- 3 files changed, 3 insertions(+), 2 deletions(-) diff --git a/share/arm.tmpl b/share/arm.tmpl index a6817f4..bf6eef9 100644 --- a/share/arm.tmpl +++ b/share/arm.tmpl @@ -12,6 +12,7 @@ kernelAddress = { 'baseline' : '0x00008000', 'highbank' : '0x00008000', 'imx' : '0x90008000', 'kirkwood' : '0x00008000', + 'mvebu' : '0x00008000', 'omap' : '0x80008000', 'tegra' : '0x00008000', } diff --git a/share/runtime-install.tmpl b/share/runtime-install.tmpl index 8e042fb..8398cde 100644 --- a/share/runtime-install.tmpl +++ b/share/runtime-install.tmpl @@ -23,7 +23,7 @@ installpkg kernel ## arch-specific packages (bootloaders etc.) %if basearch in ("arm", "armhfp"): - installpkg kernel-highbank kernel-imx kernel-kirkwood kernel-omap kernel-tegra + installpkg kernel-highbank kernel-imx kernel-kirkwood kernel-mvebu kernel-omap kernel-tegra installpkg uboot-tools %endif %if basearch == "i386": diff --git a/src/pylorax/treebuilder.py b/src/pylorax/treebuilder.py index 64ef3b9..3061f15 100644 --- a/src/pylorax/treebuilder.py +++ b/src/pylorax/treebuilder.py @@ -259,7 +259,7 @@ class TreeBuilder(object): def findkernels(root="/", kdir="boot"): # To find possible flavors, awk '/BuildKernel/ { print $4 }' kernel.spec - flavors = ('debug', 'PAE', 'PAEdebug', 'smp', 'xen', 'highbank', 'imx', 'kirkwood', 'omap', 'tegra') + flavors = ('debug', 'PAE', 'PAEdebug', 'smp', 'xen', 'highbank', 'imx', 'kirkwood', 'mvebu', 'omap', 'tegra') kre = re.compile(r"vmlinuz-(?P<version>.+?\.(?P<arch>[a-z0-9_]+)" r"(\.(?P<flavor>{0}))?)$".format("|".join(flavors))) kernels = [] -- 1.7.6.5

2 1

Custom screens for pre- and post-installation steps
by kiran.kbr＠indiatimes.com 11 Jul '12

11 Jul '12

<Resending with reformatted content to fit the viewing window> Hello, We have a kickstart-based Anaconda installer. We want to customize Anaconda to replace the console-oriented pre- and post-installation steps (i.e., %pre and %post sections from kickstart file), with their graphical counterparts. I'm new to Python, PyGTK & Glade Interface designer. Before I dig into Anaconda source code, I just wanted to check, if in theory, is it possible to do it or not. If anybody has done that, do you have any pointers/info that you would want to share on this topic. Thanks. -Kiran.

1 0

Custom screens for pre- and post-installation steps
by kiran.kbr＠indiatimes.com 11 Jul '12

11 Jul '12

Hello, We have a kickstart-based Anaconda installer. We want to customize Anaconda to replace the console-oriented pre- and post-installation steps (i.e., %pre and %post sections from kickstart file), with their graphical counterparts. I'm new to Python, PyGTK & Glade Interface designer. Before I dig into Anaconda source code, I just wanted to check, if in theory, is it possible to do it or not. If anybody has done that, do you have any pointers/info that you would want to share on this topic. Thanks. -Kiran.

1 0

[PATCH lorax/master] Actually add plymouth to the initramfs
by Will Woods 06 Jul '12

06 Jul '12

One more piece to get plymouth in place - we have to stop omitting it from the initramfs so it's available at boot. --- src/pylorax/__init__.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/pylorax/__init__.py b/src/pylorax/__init__.py index bca0ab4..cf48f0e 100644 --- a/src/pylorax/__init__.py +++ b/src/pylorax/__init__.py @@ -277,7 +277,7 @@ class Lorax(BaseLoraxClass): domacboot=domacboot, templatedir=templatedir) logger.info("rebuilding initramfs images") - dracut_args=["--xz", "--add", "anaconda pollcdrom", "--omit", "plymouth", + dracut_args=["--xz", "--add", "anaconda pollcdrom", "--install", "/.buildstamp"] treebuilder.rebuild_initrds(add_args=dracut_args) -- 1.7.10.2

3 2

Coordinating https://fedoraproject.org/wiki/Features/InitialExperience with firsboot, anaconda
by Miloslav Trmač 05 Jul '12

05 Jul '12

Hello, it seems that not all relevant parties have been talking to each other; if anyone else should be involved, please add them. In short, a new Fedora feature https://fedoraproject.org/wiki/Features/InitialExperience was proposed, replacing firstboot for the GNOME spin (only) and integrating per-system and per-user configuration. The original suggestion was for non-GNOME spins (including the DVD installation) would continue using firsboot. Now it turns out that anaconda plans to do more setup during the initial installation (including setting up the clock and adding an initial user), perhaps making all of firstboot unnecessary on non-live installations. OTOH for live-{CD,DVD} installations, the same clock/user screens would be displayed in firstboot, sharing the code; if "initial experience" plans to support firstboot screens, this (and the presumed firstboot module API changes) would affect it. Can you all talk to each other and figure out a definite plan, please? The integrated nature of IE goes against the "one installed system with multiple installed desktop environments" concept, which is sort of acceptable as long as both IE and firsboot have active maintainers, but asking the user about the same things both in anaconda and IE wouldn't do. * Which settings/screens happen in anaconda? * Which settings/screens move between anaconda and firsboot/IE (and using what mechanism)? * Which settings/screens happen both in firstboot and IE (and on which installation paths)? What code will be shared? * Which settings will be governed by each desktop environment individually? How does the transition between per-system and per-user settings happen if IE doesn't want the user to log in during the process? * Which parts of the GNOME stack will have to be installed on non-GNOME spins, or from the installation DVD when installing a non-GNOME desktop only? (and other things that I might have forgotten) Thank you, Mirek

7 6

how does anaconda determine the mountpoint of an ext3 filesystem? (during fedora upgrade from ISO)
by John Lumby 03 Jul '12

03 Jul '12

I hope it is ok to ask a question about a problem I'm having with anaconda. I am trying to upgrade a Fedora 15 system to Fedora 16 using the Fedora 16 iso, using the "upgrade existing system". This is anaconda 16.25 . I have run into the problem described in https://bugzilla.redhat.com/show_bug.cgi?id=743647 I have posted some comments there, but basically I am hitting a hard stop when anaconda comes to try to mount my F15 root filesystem, because it is failing to discover the mountpoint of this filesystem. The problem occurs about 1 second after the screen which shows the discovered existing Fedora system (which it correctly identifies as Fedora 15 on /dev/sda3), and asks me to confirm to go ahead with the upgrade. I tried some static debugging from the python source and can see that it cannot decide which is the rootDevice of the fsset because it fails to match any of the filesystem devices in the fsset to "/". But this is weird because it has already found my root filesystem and must have mounted it in order to read the fstab and so on. By the way, I have successfully upgraded a different Fedora 15 to F16 using this same iso image, so I know this iso is good. Anyway, I have two questions and hope someone can answer: . in anaconda 16.25, how does it determine what is the mountpoint of an ext3 filesystem? (pointer to relevant code appreciated) . is there any way to turn on tracing of python code during anaconda upgrade of existing Fedora? Oh - and - does anyone know if this has been fixed in later anaconda releases? John Lumby

1 0

Parsing shell-style config files
by Will Woods 02 Jul '12

02 Jul '12

So here's my shellconfig.py, presented for comment / inspiration. A couple notes: * This is an interesting construct that I wasn't aware of: def quote(s): for c in s: if c not in _safechars: break else: return s ... Apparently, if you have an 'else:' block after 'for:', it gets executed if you 'break' out of the loop. Neat! But _parseline is really the interesting part: def _parseline(self, line): s = line.strip() if '#' in s: s = s[:s.find('#')] # remove from comment to EOL s = s.strip() # and any unnecessary whitespace key, eq, val = s.partition('=') if self.read_unquote: val = unquote(val) if key != '' and eq == '=': return (key, val) else: return (None, None) * You want to remove the comment before the partition(), otherwise this line would end up with a weird value: GRAPHICS="vnc" # other choices: "spice", "none" * quote() and unquote() use pipes._safechars and shlex.split(), respectively. * pipes.quote() doesn't work because it puts single-quotes around '$', assuming that the given string is a literal filename * There's no write() method but it's pretty obvious - you'd just open a file and do f.write(str(config)) or similar. Hope that's useful to somebody, -w

2 1

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Anaconda-devel July 2012