big bootmanager / partition trouble when doing upgrade from 10.2 to 10.3
hi, i just upgraded an opensuse 10.2 (x86, download edtion) to the 10.3 goldmaster (x86, download edition). the releasenotes talk about this libata, pata, sata renaming stuff all right, but i'd never expected this complete desaster while upgrading my system. shutdown of opensuse 10.2 booting from opensuse 10.3 media (dvd iso image, burned to dvd, did media check before starting install, everything allright). first: opensuse 10.3 installer never found my previous system. the list was completely empty. so i had to check "show all partitions" or whats it called. then it came up with one opensuse 10.2 entry for my / (root) partition. allright. i had to answer some "/dev/hdaX not found" messages and was able to point to /dev/sdaX instead.... then it started to install, but at the end of the first installation step when writing the bootloader and initrd stuff, it gave me hugeloads of errors, still talking about old /dev/hdaX stuff and that it cand do all sorts of things :( so my system was completely trashed from there. trying to boot my harddisk still in the setup-mode gaves me loads of grub errors, this and that grub entries, lines and so forth not found and nothing happend. none of the grub entries worked or didn anything. it complained something about wrong entries like /dev/sda2,1 or /dev/sda2,x so i tried to boot once again from the 10.3 dvd media, and told the system to boot the installed system. to no avail. then once more, the menu "repair system"... loaded up some kernel files, and immediately resetted after that. always able to reproduce this reset-bug. whats going on with that function... is that actually supposed to work? then booted once again the 10.3 dvd, and selected install - then inside yast - other options, repair system.... to no avail again... it loads something and then gives me textmode (yast ncurses) with red error box, that some error has occured. end of story. so i tried to boot up 10.3 dvd with rescue, went into rescue login. mounted my partions into /mnt/ and /mnt/boot /mnt/var /mnt/opt and so on... tried to manually edit my /etc/fstab file, there were still loads of /dev/hdaX entries there. changed them to /dev/sdaX then i looked into /boot/ and thought that i might need to mkinitrd to be able to bootup the new system still in setupmode. i tried chroot /mnt/ but i couldnt do mkinitrd in there, as it complained about missing /dev/... and /proc/.... stuff it couldnt access. fdisk -l didnt work either, and the whole /proc/ was missing.... so i tried some final boot with the 10.3 media, selected install once more, and selected "upgrade" once again, over my partly upgraded (first step only) system. it only wanted to upgrade/install 2 more packages (some berkeley db rpms i think), then tried to write the booloater and initrd stuff all over again, and this time it succeeded. i guess because i had manually modified and fixed my /etc/fstab file in rescue mode before and the 10.3 installer also didnt complain about old /dev/hdaX entries at the beginning any more.... so now my system seems to be back online after this nightmare upgrade am i the only one to always get these install/upgrade troubles. i hate so write it again, but it has been many years back, that i didnt suffer from any upgrade problems with suse/opensuse. and i think i have pretty basic setups, no fancy disks, hardware, simple partitioning and nothing funny at all. still, suse linux manages to trash my system each and every time again, and the setup and other yast-modules crash on me on a frequent basis. :(( -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
I'm really sorry to hear about your experience. could you file a bugreport with the yast2 log files attached, please? I would have like to get this 1 month earlier now it's rather late. What kind of system do you have? The basic setups have been tested extensively, Thanks, Andreas -- Andreas Jaeger, Director Platform/openSUSE, aj@suse.de SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg) Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126
Hello, Andreas Jaeger wrote:
What kind of system do you have? The basic setups have been tested extensively,
i'm doing another cleaninstall test scenario as i write these lines. i have downloaded once more the opensuse x86 10.2 dvd iso media, burned, and installed a clean basic system (pentium3 testsystem, via chipset, /dev/hda first hdd, /dev/hdb dvd-drive). i have selected kde as desktop operating system during the 10.2 installation. after install, i did all the current opensuse 10.2 updates/patches with yast online_update right now, i have booted with the opensuse x86 10.3 dvd download dvd on this new 10.2 system: "select mode: update".... "select for update...." -> gives me once again an empty screen. show all partitions gives me the raw partitions and /dev/sda3 (i had /dev/hda1 = swap, hda2=/boot, hda3=/ and hda4=extended, inside which hda5=var, hda6=usr and hda7=opt or something like that) is being shown as opensuse 10.2, i386 architecture, ext3 filesystem. now it pops up again with the first errors: /dev/hda2 could not be mounted. i will try to give the new /hdev/sda2.... names now... it asked for the opt and for the usr mountpoints, and i changed the hda entries to sda.... it now tells me a summary of what it will update and so forth, i dont change a thing there... will see if it is able to upgrade all the packages and if the bootloader will get installed properly this time..... total of 908 packages, and 2,38 gb are being handled now... this will take a while on this box.... ..... ok so here we go again, "... save configuration" (okay...) and "install boot manager"... then error messagebox: -------------- error occurred while installing grub: grub> setup --stage2=/boot/grub/stage2 (hd0) (/dev/sda2,1) error 23: error while parsing numer: grub> quit --------------- error occured during bootloader installation. retry bootloader configuration (i will let the system stay there, until some advice of you guys, or what logfiles and stuff you'd like to have from that system) now i call this a clean basic scenario. clean x86 opensuse 10.2 from downnload dvd media. some basic partitioning with ext3, boot, /, usr, var, (home), opt upgrade to x86 opensuse 10.3 from downloaded dvd media fails utterly. cheers :( -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
Hello, on Mittwoch, 3. Oktober 2007, ab wrote: [...]
right now, i have booted with the opensuse x86 10.3 dvd download dvd on this new 10.2 system:
"select mode: update".... "select for update...." -> gives me once again an empty screen.
Reminds me to bug 299667, but I doubt your problem has the same reason. (I have /var/lib as symlink to my encrypted home, which is a quite special setup.)
show all partitions gives me the raw partitions and /dev/sda3 (i had /dev/hda1 = swap, hda2=/boot, hda3=/ and hda4=extended, inside which hda5=var, hda6=usr and hda7=opt or something like that) is being shown as opensuse 10.2, i386 architecture, ext3 filesystem.
So you have /var as separate partition. Is /var/lib/hardware populated with data?
now it pops up again with the first errors: /dev/hda2 could not be mounted.
What's the output of "mount" at this state of installation?
i will try to give the new /hdev/sda2.... names now...
Fixing /etc/fstab (and /etc/cryptotab if you use encrypted partitions) is one part. You'll also have to change the /boot/grub/device.map and maybe grub.conf (check for resume=/dev/hdaX entries). BTW: before chroot'ing to the harddisk from the recovery system, run mount --bind /dev /mnt/dev mount --bind /proc /mnt/proc mount --bind /sys /mnt/sys and you'll be able to run mkinitrd, lilo (if you use it) etc. (I learned this the hard way when a remote server didn't reboot after a kernel update.) Regards, Christian Boltz -- [Gewitterwürmchen im TFT ist verschwunden] Und ich klopfe jetzt besser nicht auf Holz - womöglich habe ich dann nachher Holzwürmer im TFT! ;-) [Dennis Kielhorn in suse-linux] -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
Christian Boltz wrote:
Hello, Reminds me to bug 299667, but I doubt your problem has the same reason. (I have /var/lib as symlink to my encrypted home, which is a quite special setup.)
i have taken a look at that bug. maybe there is a possibility about this separate /var partition at my setup. it seems (i still have that testsystem from 10.2 to 10.3 haning with the error message of the grub loader) as if on that system (still in the install-mode, very first step with booting from 10.3 dvd, and upgrading packages, but failing for grub/bootloader at the end) the old /dev/hda6 (now /dev/sda6) partition never gets mounted to /mnt/var all the other partitions are mounted during install time of yast correctly as it seems. i have found: /dev/sda3 on /mnt /dev/sda2 on /mnt/boot /dev/sda7 on /mnt/opt and /dev/sda5 on /mnt/usr when i chdir to /mnt/var i still see hugeloads of files in there.... although i suppose thats the problem. those files in var could be disjunct files from the upgrade/install process of this first opensuse 10.3 dvd media process, but they never integrated into my actual old /var (from /dev/hda6 respectively /dev/sda6) as it never got mounted. like i described in my earlier posting, at the very begining of the upgrade process of the 10.3 dvd install media, opensuse 10.3 installer doesnt find any upgradable opensuse 10.2. the list stayed empty. then i activated show all partitions, and selected the old opensuse 10.2 root partition /dev/hda3 and told the opensuse 10.3 installer that its now /dev/sda3 at that point it asked me for the locations of /opt and for /usr and i changed the old /dev/hdaX paritions to /dev/sdaX i already wondered in the back of my mind why it never asked for /var on /dev/hda6 so that i could point it to /dev/sda6.... so i guess now, thats the whole problem with the upgrade scenario.... it doesnt include the /var partition from the old clean 10.2 install. now this is a pretty big mess if you ask me. i also have checked with my previously upgraded production system (also from 10.2 to 10.3) from last night. the /var partion is properly filled with data there, but oddly the /var/log/message file stopped last night at 02:something when i did the upgrade, and messages dont arrive there any more. so i guess something nevertheless still hosed my system regarding my /var partition, even though i tried to fix it with the second upgrade process over the first upgrade step of 10.3 which then crashed with the grub/bootloader installation error. so i additionally have this problem of the /var parition and the missing logs on my production system now, any maybe i'm gonna discover even more problems with the /var partition :( damn :/ aaaargh. this sux bigtime
So you have /var as separate partition. Is /var/lib/hardware populated with data?
i have/had /boot, /, /var and /opt as separate partitons, in addition to swap. see my above explanation and hints about this bug. thanks for the hints and help with this problem. it seems as if the upgrade process fails when people have a separate /var partition and they never get asked to point to the actual /var partition in this libata-devicename-change process from /dev/hdx to /dev/sdx sigh.... -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
ab wrote:
the /var partion is properly filled with data there, but oddly the /var/log/message file stopped last night at 02:something when i did the upgrade, and messages dont arrive there any more.
i have checked on that machine, the syslog-ng conf file is still okay, like on an opensuse 10.2 before, but the syslogd hangs in the memory being defunct 1 Z root 8043 8042 0 82 0 - 0 exit 22:21 ? 00:00:00 [syslogd] <defunct> i checked another still working opensuse 10.2 system and there it reads as follows: 5 S root 4350 1 0 75 0 - 412 - Sep30 ? 00:00:00 /sbin/syslogd -a /var/lib/dhcp/dev/log -a /var/lib/named/dev/log so my productionsystems syslogd is hosed. the runlevels and the symlinks are still there, it gets started, it even creates the syslog.pid file rcsyslog status Checking for service syslog:checkproc: Empty pid file /var/run/syslogd.pid for /sbin/syslogd running so how do i fix this syslogd now is now my main question after opensuse 10.3 managed to trash my machine :( all the partitions are mounted correctly, there is empty space on all devices, the klogd runs fine and looks the same as on a normal 10.2 system. only the syslogd gets defunct :( thanks for any hints in this matter. cheers. -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
ab wrote:
only the syslogd gets defunct :(
lsof on my hosed 10.3 production system gives me this after rcsyslog start:
syslogd 10642 root cwd DIR 8,5 624 2 / syslogd 10642 root rtd DIR 8,5 624 2 / syslogd 10642 root txt REG 8,5 44048 54070 /sbin/syslogd syslogd 10642 root mem REG 8,5 1281488 161743 /lib/libc-2.6.1.so syslogd 10642 root mem REG 8,5 114636 89358 /lib/ld-2.6.1.so syslogd 10642 root 0r CHR 1,3 4542 /dev/null syslogd 10642 root 1u CHR 136,0 2 /dev/pts/0 syslogd 10642 root 2u CHR 136,0 2 /dev/pts/0
thats all. on a normal opensuse 10.2 system its rather lengthy:
syslogd 4350 root cwd DIR 8,3 520 2 / syslogd 4350 root rtd DIR 8,3 520 2 / syslogd 4350 root txt REG 8,3 44224 26762 /sbin/syslogd syslogd 4350 root mem REG 0,0 0 [heap] (stat: No such file or directory) syslogd 4350 root mem REG 8,3 1491141 5076 /lib/libc-2.5.so syslogd 4350 root mem REG 8,3 42546 5085 /lib/libnss_files-2.5.so syslogd 4350 root mem REG 8,3 129767 2530 /lib/ld-2.5.so syslogd 4350 root 0u unix 0xf7575c80 12372 /dev/log syslogd 4350 root 1w CHR 4,10 2243 /dev/tty10 syslogd 4350 root 2u FIFO 0,14 12272 /dev/xconsole syslogd 4350 root 3w REG 8,3 1457777 33032 /var/log/mail syslogd 4350 root 4w REG 8,3 1457777 33033 /var/log/mail.info syslogd 4350 root 5w REG 8,3 296862 33034 /var/log/mail.warn syslogd 4350 root 6w REG 8,3 10982 33035 /var/log/mail.err syslogd 4350 root 7w REG 8,3 0 33037 /var/log/news/news.crit syslogd 4350 root 8w REG 8,3 0 33038 /var/log/news/news.err syslogd 4350 root 9w REG 8,3 0 33039 /var/log/news/news.notice syslogd 4350 root 10w REG 8,3 1652553 44785 /var/log/warn syslogd 4350 root 11w REG 8,3 1652553 44785 /var/log/warn syslogd 4350 root 12w REG 8,3 2729179 132979 /var/log/messages syslogd 4350 root 13w REG 8,3 819225 112408 /var/log/localmessages syslogd 4350 root 14w REG 8,3 819225 112408 /var/log/localmessages syslogd 4350 root 15w REG 8,3 819225 112408 /var/log/localmessages syslogd 4350 root 16w REG 8,3 819225 112408 /var/log/localmessages syslogd 4350 root 17u unix 0xf7570580 12374 /var/lib/dhcp/dev/log syslogd 4350 root 18u unix 0xf7570200 12376 /var/lib/named/dev/log syslogd 4350 root 19u unix 0xf7575580 12378 /var/lib/ntp/dev/log
i also have tried to reinstall syslogng, syslogd and klogd packages via yast, enforce to install/reinstall the packages. also using fresh /etc/syslog-ng/syslog-ng.conf. but to no avail. my 10.3 system doesnt log any more :( do i need to upgrade-install 10.3 once again over this messed up system? maybe that would help. or do i need to cleaninstall and move my data to the fresh system? -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
Please do file a bugreport in bugzilla. For details check http://bugs.opensuse.org, Andreas -- Andreas Jaeger, Director Platform / openSUSE, aj@suse.de SUSE LINUX Products GmbH, GF: Markus Rex, HRB 16746 (AG Nürnberg) Maxfeldstr. 5, 90409 Nürnberg, Germany GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126
Andreas Jaeger wrote:
Please do file a bugreport in bugzilla. For details check http://bugs.opensuse.org,
https://bugzilla.novell.com/show_bug.cgi?id=330705 -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
Hello, on Donnerstag, 4. Oktober 2007, ab wrote:
ab wrote:
only the syslogd gets defunct :(
lsof on my hosed 10.3 production system gives me this after rcsyslog start: [nearly nothing]
I can reproduce the problem here and also have a solution, see below.
i also have tried to reinstall syslogng, syslogd and klogd packages via yast, enforce to install/reinstall the packages. also using fresh /etc/syslog-ng/syslog-ng.conf. but to no avail.
Please don't confuse syslogd (config file: /etc/syslog.conf) and syslog-ng (config file: /etc/syslog-ng/syslog-ng.conf)
my 10.3 system doesnt log any more :(
do i need to upgrade-install 10.3 once again over this messed up system? maybe that would help. or do i need to cleaninstall and move my data to the fresh system?
This bug has nothing to do with your upgrade problems. It's a bug in the syslogd apparmor profile. It seems noone tested with syslogd because syslog-ng is default since some releases. Syslogd wants to lock its pid file, which needs the extra "k" (loc_k_) permission in the latest apparmor version. Fix: Edit /etc/apparmor.d/sbin.syslogd and change /var/run/syslogd.pid lrw, to /var/run/syslogd.pid klrw, and run rcapparmor reload I just entered this in bugzilla: https://bugzilla.novell.com/show_bug.cgi?id=331444 Regards, Christian Boltz -- printk("; corrupted filesystem mounted read/write - your computer will explode within 20 seconds ... but you wanted it so!\n"); [/usr/src/linux/fs/hpfs/super.c] -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
Christian Boltz wrote:
It seems noone tested with syslogd because syslog-ng is default since some releases. Syslogd wants to lock its pid file, which needs the extra "k" (loc_k_) permission in the latest apparmor version. Fix: Edit /etc/apparmor.d/sbin.syslogd and change /var/run/syslogd.pid lrw, to /var/run/syslogd.pid klrw, and run rcapparmor reload
thankx, your fix is working all right. my logging has been restored now. the only questions is, what else on/with /var has been damaged or disabled during my 10.2 - 10.3 upgrade? kde while dragging icons to the menus, and the kde helptext seems to be trashed also. is there actually a way to reinstall (forced?) all packages, so all scripts and actions are being done again and all files are being written to the filesystems once again to be sure that everything exists, and every step/script has been executed correctly? during my second upgrade over the first trashed 10.2 one, only two more packages were being installed and the upgrade was finished after that pretty quickly. any hints on how to check if all packages work correctly? maybe with some rpm commandline parameters or something? thanks. -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
Hello, on Sonntag, 7. Oktober 2007, ab wrote:
Christian Boltz wrote: [...]
/var/run/syslogd.pid klrw, and run rcapparmor reload
thankx, your fix is working all right. my logging has been restored now.
:-)
the only questions is, what else on/with /var has been damaged or disabled during my 10.2 - 10.3 upgrade?
Good question, but I can't answer it without logging in to your system ;-)
kde while dragging icons to the menus, and the kde helptext seems to be trashed also.
Can you describe this more detailled?
is there actually a way to reinstall (forced?) all packages, so all scripts and actions are being done again and all files are being written to the filesystems once again to be sure that everything exists, and every step/script has been executed correctly?
Yes. Run Yast software management, set the filter (dropdown at the top-left of the window) to "package groups". In the list below the dropdown, scroll down and choose zzzAll. Then right-click the package list and choose "all in this list" - "force update". (back-translated from german, wording may differ)
any hints on how to check if all packages work correctly? maybe with some rpm commandline parameters or something?
rpm -Va will check all packages and list files different to the package. Be warned that many files in /var are expected to be different than the package content (otherwise they would be in /usr ;-) , so not every listed file in /var is broken by the failed update. Regards, Christian Boltz --
Runden kann das aber offenbar nicht. 18/913 sind naemlich ~1.97%. Na ja, es soll ja auch kein Statistik-Programm sein. (^-^) Sicher? Oder hast du schon mal Statistiken gesehen, die stimmen? *fg* [>> David Haller, > Sandy Drobic und Michael Raab in suse-linux] -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
Christian Boltz wrote:
kde while dragging icons to the menus, and the kde helptext seems to be trashed also.
when i search for help inside kde programs (help button, about box, or the help mouse-icon) the help subsystem comes up with loads of errors that this and that help file/chapter cannot be found and so on. maybe the system didnt install all the help packages, dont know. the dragging/icon stuff: i clicked on firefox (clean kde) and it added itself after a while to the favorites tab if i am not mistaken. there i tried to rightclick it and told kde to "add item to main panel". the firefox icon appeared but left of the suse geeko icon, and couldnt be dragged to the right where the other icons (using rootuser) like console, yast, konqueror and desktop access are located... it jumped around back and forth, but most of the time it rearranged itself to the left of the geeko. and after a while there was a blue cogwheel symbol inside my favorites tab too, which said run/sbinfirfox or something, abut never did a thing, and rightclicking brought up something totally different or similar stuff like that. somehow the icons represented something totally different than they were linked to files in the filesystem. it seems quite a mess. :( Oh and something about the graphics driver: i have some nvidia nforce6 (integrated) chipset/mainboard with that system, so i installed the nvidia driver/repostitory from the opensuse.org website. it installed the nvidia-default and so forth packages, but the "systeminformation" window tells me that i have and nvidia gfx card, but unknown what driver exactly and if 3daccelleration or not. although the windows and scrolling moves just fine and fast since the installation of that driver. so this messed up system cant determine the proper state of the nvidia driver it seems. on a clean 10.3 install on a very same box, or directly 10.2 install with sata-only drives and upgrade to 10.3 from there, and installation of the same nvidia drivers, the system reports proper nvidia drivers and 3d accel turned on. on the messed system i have: #lsmod | grep -i nv nvidia 6242424 0 agpgart 35764 1 nvidia # rpm -aq | grep -i nv x11-video-nvidiaG01-100.14.19-4 nvidia-gfxG01-kmp-default-100.14.19_2.6.22.5_30-1.1 so the drivers and modules are properly installed and also somehow working, only the systemtools doesnt really detect it or something else is still wrong.
Yes. Run Yast software management, set the filter (dropdown at the top-left of the window) to "package groups". In the list below the dropdown, scroll down and choose zzzAll. Then right-click the package list and choose "all in this list" - "force update".
will try to fore-install the packages when i get back to the box. thanks for the hint.
rpm -Va will check all packages and list files different to the package. Be warned that many files in /var are expected to be different than the package content (otherwise they would be in /usr ;-) , so not every listed file in /var is broken by the failed update.
what are those outputlines of the -Va parameter roughly? c means that this file has been changed or something? and at the beginning of each line there is something like ....L... .....UG. S.5....T and stuff like that. it reports me also missing stuff like this: missing /usr/lib/libjasper-1.701.so.1 missing /usr/share/applications/MozillaFirefox.desktop (on the clean 10.3 system or sata only systems the firefox icon sits there properly) missing /usr/lib/libsane-dll.so.1 missing /var/lib/YaST2/cd_update.desktop Thanks & regards. -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
ab wrote:
Oh and something about the graphics driver: i have some nvidia nforce6 (integrated) chipset/mainboard with that system, so i installed the nvidia driver/repostitory from the opensuse.org website. it installed the nvidia-default and so forth packages, but the "systeminformation" window tells me that i have and nvidia gfx card, but unknown what driver exactly and if 3daccelleration or not. although the windows and scrolling moves just fine and fast since the installation of that driver. so this messed up system cant determine the proper state of the nvidia driver it seems.
ok, so one more thing to the mess with the graphics driver / 3daccel: today i tried to test some compiz stuff, only it doesnt work: my Xorg.93.log tells at the end:
(EE) Failed to initialize GLX extension (Compatible NVIDIA X driver not found)
but as i wrote before, the nvidia drivers are installed and are also loaded and working (the graphics speed is much higher than before the nvidia driver installation right after the the 10.2 to 10.3 upgrade procedure (i didnt use the nvidia drivers on my 10.2 system before, just tried to test the proprietary nvidia gfx drivers on 10.3 the very first time on this system) so my system gives me: # rpm -aq | grep -i nvi nvidia-gfxG01-kmp-default-100.14.19_2.6.22.5_30-1.1 x11-video-nvidiaG01-100.14.19-4 # lsof | grep -i nvi Xorg 7894 root mem REG 8,5 2324 259410 /usr/lib/tls/libnvidia-tls.so.100.14.19 kdesktop_ 8123 root mem REG 8,5 2324 259410 /usr/lib/tls/libnvidia-tls.so.100.14.19 # lsmod | grep -i nvi nvidia 6242424 0 agpgart 35764 1 nvidia and # hwinfo --gfxcard (also the graphics version of it, when started from inside kde) gives me: Vendor: pci 0x10de "nVidia Corporation" Device: pci 0x0240 "GeForce 6150" ... Driver: "nvidia" Driver Modules: "nvidia" ... Driver Info #0: XFree86 v4 Server Module: nvidia Driver Info #1: XFree86 v4 Server Module: nvidia 3D Support: yes Config Status: cfg=no, avail=yes, need=no, active=unknown Maybe there is this problem with the active=unknown? I dont have a clean (10.2 ->) 10.3 system here right now to test what all this looks like on a non-messed-up system. I followed the rules on http://en.opensuse.org/Compiz_Fusion and http://en.opensuse.org/Using_Xgl_on_SUSE_Linux Any hints? Maybe its really time to cleanly reinstall this f*cked-up 10.2-to-10.3upgraded system after all :(((((( Thanks. -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
Hello, on Montag, 8. Oktober 2007, ab wrote:
what are those outputlines of the -Va parameter roughly? c means that this file has been changed or something?
If you are talking about the (often empty) second column: It means: (from man rpm) c %config configuration file. d %doc documentation file. g %ghost file (i.e. the file contents are not included in the package payload). l %license license file. r %readme readme file.
and at the beginning of each line there is something like
....L... .....UG. S.5....T and stuff like that.
The first column shows the status of a file. From man rpm again: Each of the 8 characters denotes the result of a comparison of attribute(s) of the file to the value of those attribute(s) recorded in the database. A single "." (period) means the test passed, while a single "?" (question mark) indicates the test could not be performed (e.g. file permissions prevent reading). Otherwise, the (mnemonically emBoldened) character denotes failure of the corresponding --verify test: S file Size differs M Mode differs (includes permissions and file type) 5 MD5 sum differs D Device major/minor number mismatch L readLink(2) path mismatch U User ownership differs G Group ownership differs T mTime differs
it reports me also missing stuff like this: missing /usr/lib/libjasper-1.701.so.1 missing /usr/share/applications/MozillaFirefox.desktop (on the clean 10.3 system or sata only systems the firefox icon sits there properly) missing /usr/lib/libsane-dll.so.1 missing /var/lib/YaST2/cd_update.desktop
Sounds like there are broken packages. Did you try the suggested "reinstall all" using YaST as described in my last mail? Regards, Christian Boltz -- Wenn der IE ein Browser ist, dann produziert Frontpage auch HTML ;-)) [Heinz W. Pahlke in suse-linux] -- To unsubscribe, e-mail: opensuse-testing+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-testing+help@opensuse.org
participants (3)
-
ab
-
Andreas Jaeger
-
Christian Boltz