Anaconda-devel May 2008

anaconda-devel@lists.stg.fedoraproject.org

38 participants
27 discussions

[PATCH] removal of libdhcp
by David Cantrell 21 Dec '08

21 Dec '08

No patch attached because it's quite large. I apologize for the length, but considering this replaces a library and introduces a new struct it meant a new file and changing calls everywhere. http://dcantrel.fedorapeople.org/anaconda/iface/ iface.h - Header file explaining the iface API. iface_t structure replacing pumpNetIntf and networkDeviceConfig, and all of the functions. iface.patch - The changes, it's about 90% complete at this point. *.txt - Explanations and notes. This patch removes anaconda's use of libdhcp, which also means removing libdhcp4client and libdhcp6client. For static network configuration, I am using libnl. To gather current network interface information, I am using libnl. To control IPv6 autoconf, I read/ write /proc since that's all we can do right now. For DHCP and DHCPv6, I run dhclient and dhcp6c, respectively. I'd like everyone to have a look at iface.h and the patch file. It's not complete yet, so I probably know about the obvious things (the FIXMEs and the useless debugging printfs and the incomplete isys.py code and so on). The goal with iface.patch is to get us closer to using NetworkManager in stage 1 and stage 2. I envision the NM changeover to be just as large, but hopefully by then we will have decided to nuke certain parts of loader entirely. Completely eliminating libdhcp and friends will be a nice step. There are some other things that need discussion as well. mkinitrd currently uses libdhcp, for example. This had me thinking we could generate libisys in an anaconda-devel package for things like mkinitrd, but I'm just brainstorming. Maybe I should put the iface code elsewhere? I dunno. At any rate, mkinitrd needs to be not left out. Work in progress, comments welcome. -- David Cantrell <dcantrell(a)redhat.com> Red Hat / Honolulu, HI

8 28

Hebrew support for anaconda
by Oron Peled 01 Jun '08

01 Jun '08

Hi, I'm the Hebrew translation team coordinator (http://translate.fedoraproject.org/languages/he) I just opened a bug (including the minimal needed patches) to add Hebrew support to anaconda (https://bugzilla.redhat.com/show_bug.cgi?id=449238) As soon as this support goes in, we can do a proper translation of anaconda to Hebrew and have it in time for F10. BTW: What is the easiest way to test a modified anaconda? For these patches, I modified my working files in /usr/lib/anaconda and tested by: anaconda -t -G -m foobar However, this runs only as root, and help me only in the first few screens (before the partitioning phase). Any better way? Thanks, -- Oron Peled Voice/Fax: +972-4-8228492 oron(a)actcom.co.il http://www.actcom.co.il/~oron ICQ UIN: 16527398 "If you have an apple and I have an apple and we exchange apples then you and I will still each have one apple. But if you have an idea and I have an idea and we exchange these ideas, then each of us will have two ideas." -- George Bernard Shaw

2 1

Customized Distribution
by Ali Majdzadeh 31 May '08

31 May '08

Hello All Following the instructions mentioned at the following address: http://fedoraproject.org/wiki/Anaconda/Customization I am trying to build a custom distribution based on CentOs 5.01. I have performed the following steps: $ mount /dev/cdrom /mnt $ mkdir ~/mydistro $ cp -avp /mnt/* ~/mydistro $ mv ~/mydistro/CentOS ~/mydistro/mydistro $ cp /mnt/.discinfo ~/mydistro $ createrepo -g ~/mydistro/repodata/comps.xml ~/mydistro/mydistro $ mkdir -p ~/mydistro/mydistro/base $ cp -avp ~/mydistro/repodata/* ~/mydistro/mydistro/base $ buildinstall --version 2 --product "MyDistro" --release "MyDistro Linux" --prodpath mydistro ~/mydistro $ rm -Rf ~/mydistro/.olddata $ pkgorder ~/mydistro i386 mydistro | tee ~/mydistro/pkgfile In the last step, i get the following error: kernel-xen-2.6.18-53.el5.i686.rpm kernel-PAE-devel-2.6.18-53.el5.i686.rpm kernel-debug-devel-2.6.18-53.el5.i686.rpm kernel-PAE-2.6.18-53.el5.i686.rpm kernel-xen-devel-2.6.18-53.el5.i686.rpm kernel-debug-2.6.18-53.el5.i686.rpm kernel-2.6.18-53.el5.i686.rpm kernel-doc-2.6.18-53.el5.noarch.rpm kernel-headers-2.6.18-53.el5.i386.rpm kernel-devel-2.6.18-53.el5.i686.rpm Traceback (most recent call last): File "/usr/lib/anaconda-runtime/pkgorder", line 164, in ? addGroups(ds, ["core", "base", "text-internet"]) File "/usr/lib/anaconda-runtime/pkgorder", line 104, in addGroups processTransaction(ds) File "/usr/lib/anaconda-runtime/pkgorder", line 76, in processTransaction ds.populateTs(keepold=0) File "/usr/lib/python2.4/site-packages/yum/depsolve.py", line 150, in populateTs self.downloadHeader(txmbr.po) File "/usr/lib/python2.4/site-packages/yum/__init__.py", line 774, in downloadHeader raise Errors.RepoError, saved_repo_error yum.Errors.RepoError: failure: CentOS/kudzu-1.2.57.1.15-1.el5.centos.i386.rpm from anaconda: [Errno 256] No more mirrors to try. Any ideas? Why does it try to check the yum mirrors? Is there anyway to perform the operation locally? Thanks in advance. Kind Regards Ali Majdzadeh Kohbanani

5 11

Anaconda Wiki
by Greg Morgan 30 May '08

30 May '08

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 With one demanding day job, three dogs, three cats, one wife, three kids and all the other kids that spawns at the house, I haven't had time to contribute to any of my favorite open source projects. I did find time to work on the initial Moin to Mediawiki conversion earlier in the week. ~ I don't know that I really changed the content of the first page, https://fedoraproject.org/wiki/Anaconda, but you want to review it--hint hint; nudge nudge; wink wink. Chuckle...that Jeremy Katz is quite the artist as noted here https://fedoraproject.org/wiki/Image:Anaconda.png. I used his work to make a banner icon so that the Anaconda pages can go toe to toe with the other project's pages (tongue in check and grinning). Jeremy's banner icon really adds _character_ to the site (subtle bad pun intended Hint: view the page source at http://people.redhat.com/~katzj/anaconda/) However, I was wondering how to go about getting Fedora Art to add that splashy reflection stuff to the banner logo? Perhaps they can embellish the banner logo altogether. Besides adding {{header|anaconda}} at the top of each Anaconda page, https://fedoraproject.org/wiki/Category:Anaconda, and a few other Moin to Mediawiki conversions, the content needs to be updated. I'll perform some mindless regexing and fix the conversion niggles but I am not sure that much of the existing content reflects how Anaconda does business, currently. Remember its just a wiki. Y'all can jump in on the action by seeding new pages, adding ideas to talk pages, adding bullet points of what needs to be covered. As always thanks for making Fedora accessible for direct contributions! Regards. Greg -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iD8DBQFIP+56xyxe5L6mr7IRAqnkAJ9W05TlDgUFBEXEvG0kO8TbmdWluACfSoNs /DMFlF/LStNrNx+HonRsdyk= =5EgI -----END PGP SIGNATURE-----

3 2

file bugs to bugzilla from anaconda
by Chris Lumens 29 May '08

29 May '08

Here's an exciting and perhaps somewhat controversial patch set. This one removes support for saving bugs via scp and adds in its place support for saving directly to bugzilla. High points: - Uses the product.bugUrl, with some mangling and checking. - Requires the user to already have an account. - Prompts for a summary so we don't have to make up something dumb. Figures out all the other bug fields automatically. - Attaches the complete traceback message. We can add whatever other files we want later on. - Attempts to detect duplicate bugs. I do this by concatenating the file name, function name, and text from each stack frame together and then hashing that. The hash is stored in the status_whiteboard. When I go to make a new bug, I first look for any bugs with the hash present. - When duplicates are found, the traceback is added as a comment and the user is put on the CC list. - Tries to bring up the network first if none is active. What still needs to be done: - Error handling needs a little more testing. - Right now I search for bugs where the status_whiteboard IS the hash. I need to be searching for bugs where the status_whiteboard CONTAINS the hash. I have ideas for how to do this from Will. So, what does everyone think? Personally I'm a big fan of this because it removes some back-and-forth on getting bug information and doesn't require the user to have another computer set up and ready to scp things to. - Chris

9 22

[PATCH] use repo config files instead of a list
by Chris Lumens 29 May '08

29 May '08

In past releases of Fedora, we've maintained a list in the install class that specifies extra repos that should be added to the task list UI. In the last release (F9), we got nailed because the mirror URLs changed and we didn't notice, so clicking "Additional Fedora Software" failed. Further, people keep wanting the updates repo enabled. What better way to fix up both these problems than to use the existing /etc/yum.repos.d/* files from the fedora-release package instead of maintaing our own list? That's what the attached patch does. It's not complete yet but it is getting there. Problems: - Driver disks should write out repo files as well, but I think driver disks are broken in general so it's not really worth fixing this immediately. - The repo configured in loader that gets passed to anaconda as method= is enabled, as is the regular Fedora release repo from the fedora-release package. This sounds like an exciting battle. I suppose it can be fixed a little bit with cost=, but I probably need a way to only enable one "base" repo. Thoughts? - Chris

7 9

There is an issue during "anaconda --test"
by zchen 28 May '08

28 May '08

I am building a latest CentOS5 and every newest stuff is ready. All are in directory "/linux/CentOS/5-respin/os/x86_64". My OS is CentOS5 too. I want to add some other pictures during installation process, so anaconda adjustment is needed. When I run command "anaconda --test --method=hd://linux/CentOS/5-respin/os/x86_64", a error message followed: Traceback (most recent call last): File "/usr/sbin/anaconda", line 702, in ? for i in os.listdir("/mnt/runtime/usr/sbin"): OSError: [Errno 2] There is no such file or directory: '/mnt/runtime/usr/sbin' Which option should I added? Did I forgot something else? Is there anyway to fix it? Thanks in advance.

1 0

odd error.
by Matt Rose 22 May '08

22 May '08

As part of an anaconda install, I'm running jetty, and feeding data into a postgresql database through jetty using another java program. (Don't ask why we're doing this. That way lies madness) This command works fine outside of anaconda, but when I try it as part of the CD installation using anaconda, Java crashes with sig11 on the same malloc and memset call each time. >From my rudimentary C programming, a signal 11 is caused by reading unassigned RAM, or RAM that's somehow faulty. Given that this problem seems to only occur in anaconda, what in anaconda could be causing java to try to access this memory address that causes it to crash? If it helps, I'm using anaconda-11.2.87 Stack: [0x90538000,0x905b9000], sp=0x905b6d54, free space=507k Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code) C [libc.so.6+0x6e077] memset+0x37 C [libc.so.6+0x6812e] __libc_malloc+0x7e V [libjvm.so+0x4f7c4a] V [libjvm.so+0x198705] V [libjvm.so+0x198c26] V [libjvm.so+0x344492] V [libjvm.so+0x255f58] V [libjvm.so+0x2a29ad] V [libjvm.so+0x29f830] V [libjvm.so+0x247149] V [libjvm.so+0x2a6d1a] V [libjvm.so+0x2a6726] V [libjvm.so+0x5b5d7d] V [libjvm.so+0x4fde19] C [libpthread.so.0+0x543b] -- Matt Rose <mrose(a)n-able.com>

3 4

bug #374891
by Adam Pribyl 22 May '08

22 May '08

I'd like to bring attention of anaconda developers to bug #374891 https://bugzilla.redhat.com/show_bug.cgi?id=374891 This is a long term bug in anaconda causing crash at update. Tim Liim does there a very good analysis and proposes solution to fix this. It's pity that F9 is still affected, but I am looking forward to F10 to fix this. Regards Adam Pribyl

1 0

[PATCH] Don't delete cached packages with %posttrans
by Will Woods 20 May '08

20 May '08

Given the following: 1) Anaconda removes packages from the cache after they are installed. 2) Anaconda runs package scripts by pulling them from the package files. 3) PreUpgrade works like an HTTP install where all the packages are already cached. Guess what happens when we try to run the %posttrans script in the kernel RPM? Yep: we've deleted the kernel package, so then anaconda prompts you to bring up the network so it can download it again to get the script. The attached patch *should* make anaconda skip package deletion *iff* the package has a %posttrans script. Two caveats: 1) I'm not sure where the cache lives during regular HTTP installs. If it's in memory, we probably don't want to leave the kernel package sitting in RAM for the rest of the install. So this patch might need to be modified so the condition is: if flags.cmdline.has_key("preupgrade") and hdr[rpm.RPMTAG_POSTTRANS]: but I'm not sure. 2) I wish I could say I've tested this patch but I've botched three attempts in the past three hours. I'm sending it to the list anyway, to get some discussion and inspection and I'll try testing it again tomorrow. -w (ps thanks for helping me figure this out, Jeremy)

3 2

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Anaconda-devel May 2008