I am attempting to install x86 gentoo on a brand new amd64 system (three days and counting) I loaded the latest console installer using the script from the FAQ. The system collects my input and then goes off to do the 28 steps it wants for install. It hit an error emerging a pci-2.20-r1 package and offered to try again. I hit try again and the system starts over from step 1, instead of 13 or so that it is on. The problem comes in when it tries to mount /proc, which is already mounted. It errors out and offers to try again, whereupon it starts over, generating the same error. The graphical install on the 2006.1 live CD will not let you continue (or backup) if you do not click to verify the root password. Your only option is to save your configuration and exit. Frustration level is very high. Reproducible: Always Steps to Reproduce: 1. start the install 2. The initial error is an emerge error, it can't find some package. The package changes each time. But I get an emerge error every time too. 3. If you try to pick up where you left off, it starts over at step 1 of 28. Actual Results: Infinite install loop. Install will never finish! Expected Results: I would like the ability to ignore the mount /proc fs, because I know it is already mounted. I really want the install to pickup on the same step that it errored out on instead of starting over, including copying all the packages to the drive. LI: March 29 2007 18:23:46 - Edited Config file /mnt/gentoo/etc/conf.d/clock GLI: March 29 2007 18:23:46 - Config files updated using etc-update. make.conf$GLI: March 29 2007 18:23:47 - Edited Config file /mnt/gentoo/etc/make.conf GLI: March 29 2007 18:23:47 - Edited Config file /mnt/gentoo/etc/make.conf GLI: March 29 2007 18:23:47 - Make.conf configured GLI: March 29 2007 18:23:48 - Exception received during 'Preparing chroot': Mou$GLI: March 29 2007 18:23:48 - Traceback (most recent call last): GLI: March 29 2007 18:23:48 - File "../../GLIClientController.py", line 122, in$ func() GLI: March 29 2007 18:23:48 - File "../../GLIArchitectureTemplate.py", line 542ne 542, in prepare_chroot $ raise GLIException("MountError", 'fatal','prepare_chroot','Could not mount $mount /proc') GLI: March 29 2007 18:23:48 - GLIException: MountError :FATAL: prepare_chroot: root: Could not mount /proc $ <?xml version="1.0" ?> <gli-profile> <kernel-build-method> genkernel </kernel-build-method> <cron-daemon> vixie-cron </cron-daemon> <verbose> False </verbose> <domainname> wetzker.org </domainname> <portage-snapshot> file:///mnt/cdrom/snapshots/portage-2006.1.tar.bz2 </portage-snapshot> <kernel-source> livecd-kernel </kernel-source> <install-pcmcia-cs> False </install-pcmcia-cs> <time-zone> America/Chicago </time-zone> <install-distcc> True </install-distcc> <boot-device> /dev/hda </boot-device> <root-mount-point> /mnt/gentoo </root-mount-point> <dynamic-stage3> True </dynamic-stage3> <kernel-bootsplash> False </kernel-bootsplash> <bootloader-mbr> True </bootloader-mbr> <hostname> awetzker </hostname> <bootloader-kernel-args/> <recommended-partitioning/> <post-install-script-uri/> <nisdomainname/> <install-stage> 3 </install-stage> <ftp-proxy/> <stage-tarball/> <bootloader> grub </bootloader> <root-pass-hash> $1$CI7hA.5I$tz679cb/gH5Up.JBOsuM50 </root-pass-hash> <http-proxy/> <rsync-proxy/> <grp-install> True </grp-install> <logging-daemon> syslog-ng </logging-daemon> <kernel-config/> <portage-tree-sync> snapshot </portage-tree-sync> <kernel-initrd> True </kernel-initrd> <yes-iam-sure> False </yes-iam-sure> <mta/> <default-gateway interface="eth0"> 192.168.2.1 </default-gateway> <dns-servers> 68.87.77.130 68.87.72.130 </dns-servers> <install-packages> gnome xorg-x11 kde-meta chkrootkit nmap crack-attack xpdf netcat samba postgresql apache keychain gvim ufed rdesktop vim ethereal ntp koffice xmms xine-ui gimp mozilla-thunderbird openoffice-bin mozilla-firefox mplayer gaim k3b </install-packages> <mounts/> <network-interfaces> <device broadcast="192.168.2.255" ip="192.168.2.39" netmask="255.255.255.0"> eth0 </device> </network-interfaces> <network-mounts/> <services> alsasound,apache2,distccd,hdparm,sshd,xfs,xdm </services> <users> <user comment="" groups="users,wheel,audio,games,cdrom,video" hash="$1$BkwmyrtM$OYHR3VVUsB3kyNt.nyUFt." homedir="/home/almaw" shell="/bin/bash"> almaw </user> </users> <etc-files> <file name="rc.conf"> <entry name="DISPLAYMANAGER"> gdm </entry> <entry name="XSESSION"> kdm </entry> </file> <file name="make.conf"> <entry name="DISTCC_HOSTS"> localhost </entry> <entry name="FEATURES"> distcc </entry> </file> <file name="conf.d/clock"> <entry name="CLOCK"> UTC </entry> </file> </etc-files> </gli-profile> livecd home #
I am sorry to hear about your frustration with GLI. The installer is designed to follow a set list of procedures in order to install. Each of those steps depends on the other being executed. There is currently no resuming support, though there is retry support in the console frontend for 2007.0, which is being released in a few weeks. To clarify your bug, you are complaining about the installer always starting at step 1 instead of being able to pick a point to resume an install. This isn't possible due to the fact that the installer needs to know things like the mountpoints and which way portage was sync'd in order to know how to do other steps down the line. I would recommend waiting a few weeks for 2007.0 to come out if you can. Otherwise you may be able to avoid your problem by not using the snapshot on the livecd and downloading your stage3 and portage from the internet.
To clarify, it looks like you loaded a previously created install profile, which triggers a special mode in the new version. This special mode runs all the steps at once instead of in small groups. When the installer asks if you want to retry, it will start back at the beginning of that small group of steps. In this case, that small group of steps was the entire install. It's not so much a bug as a lack of a feature (more fine-grained retrying).
