This bug is for tracking bugs on the test-livecds. It is used to help the livecd-team to keep track of all the stuff that should be / needs to be fixed. The test-livecds have a datestamp in their filename - please mention this datestamp when reporting anything.
livecd 20050110: - bug with some messages about inability to remove some files on bootup (initrd) - splash can't start since local wasn't started (origins from a wrong /etc/conf.d/local.start)
Issue #1 should be fixed by a patch I am currently testing to genkernel. Issue #2 should be resolved now in catalyst.
(Minor) bugs I have with the 20050110 LiveCD: - Is the error message Booting.rm: cannot remove `/tmp/.initrd/dev/fd': No such file or directory normal? The system seems not to be harmed by that. - vim warns about "filetype plugin on" in line 62 of the file /etc/vim/vimrc - less gives a warning about /usr/bin/lesspipe.sh: line 1: file: command not found
New test-iso: x86-installcd-minimal-2005.0-20050115.iso FIXED: - vim error - less error REMAINING: - bootsplash-weirdness - "rm"-messages on bootup
I just tested the 20050115 ISO: - the keyboard chooser does not show up. It just starts with a us layout - less shows a warning with "file: could not find any magic files!" but it does not harm anything.
file: /usr/share/misc/file gets removed on livecd-stage2 - will put it back and see how much space it takes the keyboard-selector came up in vmware and on my thinkpad here. If you pressed a key (as we talked about in irc), try if it works without pressing a key on bootup.
Yep, the layout-selector comes up, but scrolled quite fast away. The timeout is IMHO a little bit short.
The timeout is 10 seconds and stops on any keypress. This is the same as 2004.3, and we didn't seem to have any complaints on it then.
new test-iso: 20050119 Fixed: - "file"-bug - added missing prism54-firmware (/lib/firmware/isl3890) Remaining: - rm-messages on bootup - bootsplash
fixed: - bootsplash - rm-messages (finally ;) ) new: - when evms is enabled, I get a complaint about /lib/evms/2.4.1/swap-1.1.11.so failing to load. This file, however, exists in that location.
I have several problems: 1. Shutdown cannot shutdown properly with x86-installcd-minimal-2005.0-20050119. Just type at console directly after boot: $ poweroff Last message printed on screen is "Remounting remaining filesystems readonly", the power for the hard drives is turned off and then nothing else. CPU stays active as well as the screen. I have tested this on two different computers. One system is Athlon Thunderbird based on a Asus A7N8X Deluxe (nForce 2), the other one is a Pentium 166 on a Asus P55T2P4. 2. Use boot parameter acpi=on Error message: /etc/init.d/acpid: no such file
With the http://gentoo.osuosl.org/experimental/x86/livecd/x86/ minimal CD dated 20050125. I can't mount any partition because it says either that partition is already mounted or the mount point is busy. So I try a different partition and same result. I make different mount points and same result. I try to umount the partitions and it says they're already unmounted. Looking at /etc/mtab, it doesn't show any drives mounted except for the tmpfs and loopbacks by the liveCD. Any more info... just let me know.
Confirmed. However it's strange cos this didn't happen with one of the former isos (I pulled a small system up in a virtual machine). Examining.
Poweroff problem and missing acpid confirmed for 2005.0-rc1. Have also the mount problem. Additionally I cannot boot from my 2nd CD-ROM, boot process fails after looking for the image at /newroot/dev/sr0, but it does not check /newroot/dev/sr1 (Hint: my DVD at /sev/sr0 has problems reading some CD-RWs, so I used the writer to boot).
BTW: What's the policy for the available utilities on the stage1? It would be nice to have "dar" to restore from backups.
The policy for the contents of the stage-files are: stage1: only what is needed as a very minimal toolchain to be able to run "emerge --sync" and /usr/portage/scripts/bootstrap.sh stage2: whatever results from the bootstrap from the stage1 without setting any useflags (except those set via the profile) stage3: whatever results from "emerge system" without setting any useflags (except those set via the profile)
The policy is: absolutely nothing but that required to install Gentoo
New iso: RC4 Bugs fixed: - acpi not working - direct mounting of devices impossible - shutdown won't work What needs to be tested: ACPI now is OFF by default. What we need to know is: - Does ACPI + Hyperthreading work (on appropriate machines) with "acpi=ht" on the bootprompt? - Does ACPI work properly with "acpi=on" on the bootprompt? Please ignore the GPM-Message on bootup. It's just cosmetic, your mouse will work (you can consider this "bug" as being fixed).
Testing on a HP Compaq nx9000 Laptop. 2.2GHz P4m inside. > - Does ACPI + Hyperthreading work (on appropriate machines) > with "acpi=ht" on the bootprompt? It boots, but ACPI isn't working. Although /proc/cpuinfo has a ht-flag, the machine isn't ht-capable (stupid P4m...). > - Does ACPI work properly with "acpi=on" on the bootprompt? Works fine. No problems!
acpi=on works fine now with rc4. Can't say something about HT, do not ave such a beast. Poweroff is not completly fixed! It still fails on my old Pentium system on the ASUS P55T2P4 (as described above), while it works now for my Athlon system.
The old pentium does not have acpi. You will be required to use acpi=off doapm to get it to power off.
I have a new rc ready that should be out on the mirrors soon. http://gentoo.osuosl.org/experimental/x86/livecd/x86/ The image is named: x86-installcd-minimal-2005.0-rc5.iso This is considered to be the final shot, so it should be quite identical to what is released as final. Bugs fixed: - acpi not working - gpm-error-message
rc5 good on an ibm x225 single xeon 2.66 (ht) with default options (none), acpi=on and acpi=ht. In the 3 cases cpuinfo shows "two" processors.
> The old pentium does not have acpi. You will be required to use acpi=off doapm to get it to power off. acpi is off by default and APM is found by the autoconfigure, but just to be sure I booted with acpi=off and doapm and it was no difference. rc5 is different though, poweroff will now shutdown the hard drives and I cannot switch between the terminals anymore, but the CPU is still not shut down. What's supposed to happen with "nousb"? I recognized that the different USB modules are not tried, but the system ist still says "Mounting USB device filesystems (usbfs)" and "Coldplugging usb devices". /newroot/dev/sr1 is not checked for the boot image fbsplash still displays 2004.3 in the bottom
Using 2005.0-rc5 no network (e100 not loaded) on system using an Intel 915 board (onboard lan) and a firewire device unless the nofirewire option is provided (but then there is no 1394 support); 2004.3 boots normally and loads both the 1394 and the e100 modules.
Incoming from the peanut gallery: any chance of getting the n_hdlc driver added to the .config for the livecd kernel? It's the only thing that's stopping me getting my SpeedTouch 330 ADSL modem working with the userland tools on the livecd. :-)
Looks like a typo sneaked into the linuxrc (using x86-installcd-minimal-2005.0-rc5.iso): if [ "${DO_evms2}" -a "${USE_LVM2_NORMAL}" -eq '1' ] should be: if [ "${DO_evms}" -a "${USE_EVMS_NORMAL}" -eq '1' ] Oh, and isolinux/help.msg needs to be updated to reflect the new boot options.
Created attachment 51348 [details, diff] Patch againt livecd-functions.sh This patch correctly exports CDBOOT even if using a cdroot=/dev/foo parameter.
Thanks for the reports... I'll be sure to get this stuff in as soon as I get a chance... plasmaroo: can you get a fixed genkernel out for me?
Is it possible to compile the driver for the Tekram DC395 SCSI controller in the liveCd kernel? I need it to boot because I have only SCSI cdrom :-( The config entry is CONFIG_SCSI_DC395x=y Thanks!
Not sure if this is intended or what, but if any of the install partitions are formatted as ext3, when the mount command is used it mounts as ext2 and prints a message stating such, i.e.: mke2fs -j /dev/hda3 mount /dev/hda3 /mnt/gentoo EXT2-fs warning (device hda3): ext2_fill_super: mounting ext3 filesystem as ext2 Pretty sure this isn't intended behavior, however, nothing is mentioned in the draft x86 handbook (http://www.gentoo.org/doc/en/handbook/draft/2005.0/handbook-x86.xml?part=1&chap=4#doc_chap5), so this should probably be updated, either in the livecd (because I never recall getting this message before) or by adding a -t ext3 to the docs. Thanks for the hard work guys. :)
The EVMS2 issue as reported in comment #27 is fixed in genkernel-3.1.0k.
Created attachment 52027 [details, diff] Allow a baud rate of 19200 Patch against livecd-functions.sh to allow a baud rate of 19200bps.
x86-installcd-minimal-2005.0-rc6.iso net-setup eth0 doesn't work for manual settings (no such device) but setting it up using ifconfig and route works fine. the keymap selection is hidden behind the silent splash waiting for a timeout (the user doesn't know to hit f2 to see it) if possible: with keymap turn off the splash and with nokeymap leave it on. everything else seems to work very smoothly. good job! roger55
livecd root # mount /dev/hda3 /mnt/gentoo/ EXT2-fs warning (device hda3): ext2_fill_super: mounting ext3 filesyst em as ext2 If I mount it with -t ext3 it mounts as ext3, why is that?
#34: It is documented that you must press F2 for keymap... we have no way of interrupting splash without throwing an error, so we don't want to #35: e2fsprogs bug that is only fixed in latest ~arch version that should probably be documented
x86-installcd-2005-rc9: Splash now doesn't switch to verbose mode until bootup is finished. (++) irssi doesn't work, missing libsocks... cannot mount the harddisk on my laptop, getting DMA errors, seek incomplete... the laptops harddisk works just fine with my own 2.6.10 kernel and has worked with -rc6 was there some change to the kernel config? sorry to bring the bad news. Roger
those weird dma errors go away if I poweroff my thinkpad before booting the livecd and after having booted the livecd before booting my install. with that trick it works fine. another issue: memtest86 doesn't work: invalid or corrupt kernel image
and finally a cosmetical(?) thing on shutdown: Shutting down Logical Volume Manager ... /etc/lvm/archive: mkdir failed: Read-only filesystem Failed to shut LVM down Roger
rc10: irssi works mount -t ext3 <-- still required so check if we have that documented in the on cd handbook. memtest86 works the LVM error on shutdown remains. net-setup works and my nic(s) were properly identified I can just say it works as expected. It would be nice if someone could give that one a shot with a more sophisticated setup. usb/scsi cdroms, raid, adsl etc... cheers Roger
*** Bug 85683 has been marked as a duplicate of this bug. ***
I posted bug # 85863 which has been listed as a duplicate of this bug. i think that the bug is actually not a duplicate, but if i am wrong please forgive me. the problems that i have reported with incomplete screen blanking are occurring with x86-installcd-minimal-2005.0-rc5.iso. this is the only live CD image that appears to be available on the mirrors. i have never seen a live CD image on the local mirror with a timestamp in its name. if there is a newer release candidate somewhere, i apologize for reporting something out of date, and i'd appreciate a tip about where to look for a newer ISO. i have been looking here (my local mirror): http://gentoo.netnitco.net/experimental/x86/livecd/x86/
Done.
*** Bug 90690 has been marked as a duplicate of this bug. ***
