Re: [expert] Unfixable RPM DB corruption in 8.2
On Tue, Apr 09, 2002 at 09:06:47PM -0400, A V Flinsch wrote: > I have had this happen in the past, and to avoid it in the future I do > regular backups of /var/lib/rpm Good idea for future reference, but I think there needs to be some sort of validation step added prior to backup, or it would be easy to end up with a useless backup. > What I did to fix the problem was the following (found thru lots of trial > and error) > > 1 - backup /var/lib/rpm --- just to be safe > 2 - run 'rpm -qa | sort | less ' -- take a look at the output and see if > there were any "funky" named packages -- packages named with all sorts of > odd symbols I don't see any oddly named packages, but then rpm -qa is core dumping, presumably due to the RPM DB corruption and possibly before finishing the list. Maybe you've been seeing a different sort of DB corruption. Jon Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
[expert] Unfixable RPM DB corruption in 8.2
I recently upgraded from 8.0 to 8.2 on a P3-733 machine running an Acer MB (M25D, I think). After a week or so, and without being certain of the exact sequence of events leading up to this, the RPM database seems to have become corrupted. 'rpm -q --all' lists 537 packages, ending with an old 'kernel-2.4.3-20mdk', then gets a SEGV and core dumps. Notably, I had been trying to 'rpm -e' this package at about the same time that I first noticed the corruption. There were more than 537 packages installed, as well. 'rpm --rebuilddb' cranks away for a while, then also gets a SEGV and core dumps, leaving about 18 MB of stuff behind in /var/lib/rpmrebuilddb. Both errors are repeatable. It's possible that the sequence of events leading up to this was something like: - Upgrade to 8.2 - Install and remove various additional packages (notably, removing some old packages from 8.1 and 8.2 Cookers). - At some point, fill up the partition containing /var/lib, without noticing it. - Attempt to rpm -e kernel-2.4.3-20mdk and kernel-source-2.4.3-20mdk. - Notice RPM DB corruption. However, I'm not particularly confident in stating this was the actual sequence of events. Is this known behavior? Do I have any options other than a clean re-install of 8.2? Searching through old archives of the expert list, I came across a pointer to Pascal Bourguignon's howto on rebuilding the DB: http://hermes.afaa.asso.fr/users/pascal/linux/rpm-rebuilddb I'm willing to give this a try if there's a good chance of success, though I'd first like to hear if anyone else has dealt successfully with this problem. Thanks, Jon Leech Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
[expert] Changing to security level 3 or 4 makes it impossible to change user passwords?!
[I first sent this out on the 17th, but it seems to have been lost to a power glitch at Mandrake - retrying.] I've just changed some of the security levels on one of my machines running Mandrake 8.0 - from no security whatsoever while running inside our corporate VPN, to 'msec 4' now that I'm running behind a Linksys NAT router to a DSL line. I now find it impossible to change user account passwords. I did try backing off to 'msec 3', but that didn't help. passwd claims 'all authentication tokens updated successfully' when run as root to change a user account's password (e.g. 'passwd ') - but no change occurs in the password file. In addition, /var/log/auth.log shows a message like PAM_pwdb[#]: password for (login/UID) changed by (root/0) So passwd appears to think it did something correctly - but as best I can tell, it actually does nothing at all. There are no problems changing *root*'s password, just user passwords. In addition, attempts to change user passwords when su'ed to that user simply fail: % passwd Changing password for (current) UNIX password: passwd: Authentication service cannot retrieve authentication info. Is this perhaps something to do with PAM? /etc/pam.d/passwd looks like #%PAM-1.0 auth required/lib/security/pam_pwdb.so shadow nullok account required/lib/security/pam_pwdb.so password required/lib/security/pam_cracklib.so retry=3 password required/lib/security/pam_pwdb.so use_authtok nullok shadow md5 But pam.d/passwd wasn't changed when I ran 'msec', anyway, so I'm grabbing wildly for ideas. I tried rebuilding and debugging passwd, but the problem seems to lie somewhere down inside the PAM libraries, which are harder to rebuild and understand. I also tried removing the 'shadow' keyword from the pam.d/passwd conf. file, since I'm not using shadow passwords at present, but no change in behavior resulted. Has anyone seen this behavior or know what's up? I searched in the expert mail archives and on Google, but couldn't turn up a similar problem report. Thanks, Jon Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] SGI Apache enhancements
On Tue, May 22, 2001 at 01:49:22AM +0200, Stefaans Mostert wrote: > Hallo all > > Been a while but I am back now ;-) > > I need to find the home page of SGI performance enhancemants for apache > Can anyone help me please? This is probably what you're after: http://oss.sgi.com/projects/apache/ All of SGI's open source projects are hosted on oss.sgi.com. Jon
Re: [expert] 8.0 hanging sporadically after image selection in LILO
On Tue, May 15, 2001 at 03:17:06PM -0700, Civileme wrote: > I have noticed some weirdnesses with some laptops (SONY VAIO PCG series, > especially old ones), that affects their ability to boot windows after > booting linux without being shut completely down first. > > So I want to know more about your laptop. Then maybe we can reproduce this > behavior and have a chance at finding out what is going on. It's a Gateway 9100XL (PII 266 MHz, Trident Cyber 9397/4M video, 5G/64M, 3c575TX CardBus network i/f, Telepath 33.6K CardBus modem) laptop. The model is slightly over 3 years old; I've been running various versions of RH (5.0 - 6.1) on it up until the LM 8.0 install, and have not seen this particular behavior before. What else would you like to know about it? Jon
Re: [expert] 8.0 hanging sporadically after image selection in LILO
On Mon, May 14, 2001 at 05:15:35PM -0700, Civileme wrote: > On Saturday 12 May 2001 15:15, Jon Leech wrote: > > The install went cleanly, but I'm having sporadic problems with > > booting. Some of the time LILO 21.7 comes up with the graphical splash > > screen, and selecting linux or linux-nonfb boots properly. Some of the > > time LILO comes up in text mode, and selecting linux or linux-nonfb > > hangs forever after printing 'Loading linux'. There is no apparent > > pattern in when this happens. [...] > Let's see /etc/lilo.conf > > and fdisk -l /dev/hda 'fdisk -l /dev/hda' says: Disk /dev/hda: 255 heads, 63 sectors, 623 cylinders Units = cylinders of 16065 * 512 bytes Device BootStart EndBlocks Id System /dev/hda1 * 1 261 20964516 FAT16 /dev/hda2 262 623 29077655 Extended /dev/hda5 262 300313236 83 Linux /dev/hda6 301 606 2457913+ 83 Linux /dev/hda7 607 623136521 82 Linux swap /etc/lilo.conf contains: boot=/dev/hda map=/boot/map install=/boot/boot.b default=linux keytable=/boot/us.klt lba32 prompt timeout=50 message=/boot/message-graphic menu-scheme=wb:bw:wb:bw image=/boot/vmlinuz label=linux root=/dev/hda5 append=" quiet" vga=788 read-only image=/boot/vmlinuz label=linux-nonfb root=/dev/hda5 read-only image=/boot/vmlinuz label=failsafe root=/dev/hda5 append=" failsafe" read-only other=/dev/hda1 label=dos table=/dev/hda other=/dev/fd0 label=floppy unsafe FWIW, since the initial post, my laptop seems to have evolved to a state (or, I am replicating sequences of events without realizing it, that keep it in that state) where booting from HD will freeze after 'Loading linux' in *all* cases, not just sporadically; the boot floppy seems to be the only way to get the machine up now. Jon
[expert] 8.0 hanging sporadically after image selection in LILO
(Sorry if this is a duplicate - I had a few bobbles getting subscribed, and am not sure if it went out the first time) I just installed Mandrake 8.0 on a Gateway 9100XL laptop. This used existing partitioning (2.5 /dev/hda1, Win95; 330M /dev/hda5, /root; 2G /dev/hda6, /usr; 128M /dev/hda7, swap), but was a fresh install. The install went cleanly, but I'm having sporadic problems with booting. Some of the time LILO 21.7 comes up with the graphical splash screen, and selecting linux or linux-nonfb boots properly. Some of the time LILO comes up in text mode, and selecting linux or linux-nonfb hangs forever after printing 'Loading linux'. There is no apparent pattern in when this happens. Booting from floppy always seems to work. Is this a known problem? I hunted around google and Mandrake's website for a while, but didn't turn up anything that seemed relevant. Thanks, Jon
[expert] 8.0 hanging sporadically after image selection in LILO
I just installed Mandrake 8.0 on a Gateway 9100XL laptop. This used existing partitioning (2.5 /dev/hda1, Win95; 330M /dev/hda5, /root; 2G /dev/hda6, /usr; 128M /dev/hda7, swap), but was a fresh install. The install went cleanly, but I'm having sporadic problems with booting. Some of the time LILO 21.7 comes up with the graphical splash screen, and selecting linux or linux-nonfb boots properly. Some of the time LILO comes up in text mode, and selecting linux or linux-nonfb hangs forever after printing 'Loading linux'. There is no apparent pattern in when this happens. Booting from floppy always seems to work. Is this a known problem? I hunted around google and Mandrake's website for a while, but didn't turn up anything that seemed relevant. Thanks, Jon
