Anaconda-devel January 2005

anaconda-devel@lists.stg.fedoraproject.org

29 participants
36 discussions

Start a nNew thread

[PATCH] fix exception handling in exception handlers
by Matt Wilson 07 Feb '05

07 Feb '05

I hate it when exception handlers cause unhandled exceptions. Cheers, Matt

2 1

getfullcomps.py CRITICAL ERROR
by David 31 Jan '05

31 Jan '05

When 'getfullcomps.py' returns a critical error (like "CRITICAL ERROR: Unable to find package prctl"), what should I do? I think I should find thah package and save it in the RPMS directory. Is it true? Otherwise, where does 'getfullcomps.py' looks for the packages? Other option is finding the package which requires this and delete it from the comps.xml. Is it true? Thanks. --------------------------------------------------- David Asensio Ortega (david(a)sabioss.es) Sabioss Trade S.L. Santiago 24, 2º Izq. 47001 Valladolid (España) ---------------------------------------------------

1 0

Re: where is firstboot invoked from?
by Robert P. J. Day 30 Jan '05

30 Jan '05

On Sun, 30 Jan 2005, Sp0oKeR Labs wrote: > In /etc/rc.d/rc5.d/ > I dont remember the numer exactly, bu Snumber_firstboot will run at boot... > Knumber_firstboot wont. > > Best Regards, > > If Quoting "Robert P. J. Day" <rpjday(a)mindspring.com>: > > > > > at what point in the boot process is firstboot invoked? i'm looking > > through the startup scripts, and i can certainly see the file > > /etc/rc.d/init.d/firstboot, but it's not clear who runs that. > > > > i don't see a call in /etc/rc.d/rc.sysinit, or symlinks from one of > > the rc#.d directories. so who checks for, and runs, that script? > > thanks. i've already moved this discussion to the fedora list, and i've got past this point, but there's definitely something strange with firstboot. at this point, i'm convinced there's a bug since, once it's run, there is no obvious way you can reboot and add the kernel parm "reconfig" to have it run again in reconfigure mode. anyway, it's all over there on the fedora list. rday

1 0

firstboot: never mind, found it
by Robert P. J. Day 30 Jan '05

30 Jan '05

just noticed the check for kernel parameters in the firstboot script: if grep -i reconfig /proc/cmdline >/dev/null || [ -f /etc/reconfigSys ]; then action=reconfig fi now it makes sense. by the way, the comments at the top of the /etc/rc.d/init.d/firstboot script are a bit misleading: "Firstboot is a druid style program that runs on the first time a machine is booted after install. It checks for the existence of an /etc/sysconfig/firstboot file. If it doesn't find the file, then the firstboot program needs to run. If it finds the file, firstboot will not be run." it certainly seems that that last part is not true. if /etc/sysconfig/firstboot is found, then it is further *consulted* for the value of the variable RUN_FIRSTBOOT. so that comment is overly simplistic, no? rday

1 0

where is firstboot invoked from?
by Robert P. J. Day 30 Jan '05

30 Jan '05

at what point in the boot process is firstboot invoked? i'm looking through the startup scripts, and i can certainly see the file /etc/rc.d/init.d/firstboot, but it's not clear who runs that. i don't see a call in /etc/rc.d/rc.sysinit, or symlinks from one of the rc#.d directories. so who checks for, and runs, that script? thanks. rday

1 0

RE: Diagnosing Anaconda Installation 'stalls' over a LAN
by Oden, James 29 Jan '05

29 Jan '05

> I have noticed that Anaconda may stall when installing RPMs over a > LAN connected to a hub. What sort of problems can cause Anaconda to > timeout completely? > > How can I diagnose these problems? > Not that this is your problem you are experiencing, but anaconda depends on rpm to a great degree so things that would cause rpm to stall would cause anaconda to stall. There are two "general" cases that I have seen rpm stall (though not necessarily in anaconda). One is an rpmdb lockup, the other is a scriptlet lockup (much more rarer btw). In both cases if you ran strace against rpm you would see rpm waiting on a futex (or rather it is in the FUTEX syscall). You can determine which one by running gdb against rpm and getting a backtrace. In the scriptlet lockup you have passed through runScript() in the stack at one point, with db lockups your into the db code. That said neither of these may be your lockup. Cheers...james

1 0

Diagnosing Anaconda Installation 'stalls' over a LAN
by Shawn Starr 28 Jan '05

28 Jan '05

I have noticed that Anaconda may stall when installing RPMs over a LAN connected to a hub. What sort of problems can cause Anaconda to timeout completely? How can I diagnose these problems? Thanks, Shawn.

1 0

disk initialization
by Gene Czarcinski 26 Jan '05

26 Jan '05

At one time, there was an option to run badblocks when formatting a partition during installation. Is there a reason that it was dropped? Is there some way it can be added back in? It just occurred to me that this may be available under "expert" mode. If it is, good; otherwise, could it be added to expert mode? -- Gene

2 1

Fedora Core Livecd pygtk installer, based on anaconda
by root 25 Jan '05

25 Jan '05

Hi, perhaps this is of interest: There`s a first version of an pygtk gui installer for livecd`s: The idea is to provide a gui driven wizard/interface to the process of creating livecd`s from scratch. The 'livecdanaconda installer' (which is my 'working term for it) is a clone of the fedora anaconda system installer, customized to create livecd systems from redhat/fedora installation rpms - some details are described below. I finally found (and also heard from other livecd creators that use my software) that the process to create a livecd from scratch or an existing system currently involves too much manual work and is too errorprone, so i tried to customize the anaconda installer and came uo with this cloned version tom make it installable parallel to the original anaconda. The current version is 0.8, screenshots aviable on website: http://www.linux4all.de -> livecdanaconda section and: http://www.linux4all.de/livecd/livecdanaconda If you`re interested, if have some questions: - would there be interest for a project like this in the fedora/anaconda community ? - who could help me review the installer including designing and/or programming, and give me suggestions and comments ? Many thanks in advance! Best Regards, Dirk Westfal Details of the installer: It : a) runs on an existing redhat/fedora system b) installs a livecd master system into a folder on the current system and c) is able to turn any redhat/fedora installation into a livecd. d) uses network installation source (comps,hdlists,packages) and handles: - preparation and testing of build environment - download of latest livecdrpms which contain all system modifications neccessary for a livecd - installation of a 'master' with redhat/fedora or from customized comps + hdlist files. - installation of the 'livecdrpms' to make the master livecd capable - customizing the master via konsole - creation of a virtual harddisc - loading the master to the virtual harddisc - creating a compressed copy of the virtual harddisc - creating appropiate initrd for the livecd and setup of isolinux bootloader - creation of the final live iso image The screens are still clumsy and output is done via statusbar, but at least it works and one has to start somewhere. The resulting iso images are not bootable yet - this will be fixed before 1.0, of course.

1 0

config for undetected hda
by David Becker 25 Jan '05

25 Jan '05

I am trying to use Rocks's anaconda to install on a Xen virtual machine. This fails when anaconda does not detect the existing disk partitions. I would not expect it to detect the xen virtual block device since there is no PCI or IDE bus. The hda1 and hda2 partitions exist and have the same major/minor numbers as an IDE disk. What I want to know is how to manipulate the anaconda/kickstart config so anaconda will mknod /dev/hda1 and /dev/hda2 and start writing to them even though it never detected my hda disk. The disk is partitioned and formated already. I do have this much in the kickstart config: part / --noformat --onpart hda1 part swap --noformat --onpart hda2 ..but since Anaconda thinks there are no disks, installation doesn't get past the "partitionobjinit" step in installclass.py. partitionObjectsInitialize fails at diskset.checkNoDisks(intf) because there are no disks in probedList. Can some %pre code just inform anaconda it has /dev/hda1 and /dev/hda2 even though it didn't detect anything? Or can a driver disk be hacked to do this? David Becker becker(a)cs.duke.edu

3 4

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Anaconda-devel January 2005