Red Hat Bugzilla – Bug 26935
RPM segfaults when trying to do a rpm -Uvh
Last modified: 2007-04-18 12:31:23 EDT
From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux 2.4.2-pre3 i686; en-US; m18) Gecko/20010208 When trying to do a rpm -Uvh *.rpm om a couple of packages rpm segfaults. Reproducible: Always Steps to Reproduce: 1.Just do a rpm -Uvh on a rpm package... 2. 3. Actual Results: I here do a rpm -Uvvvh D: counting packages to install D: found 1 packages D: looking for packages to download D: retrieved 0 packages D: New Header signature D: Signature size: 68 D: Signature pad : 4 D: sigsize : 72 D: Header + Archive: 616030 D: expected size : 616030 D: opening db index /var/lib/rpm/Packages rdonly mode=0x0 D: locked db index /var/lib/rpm/Packages D: opening db index /var/lib/rpm/Name rdonly mode=0x0 D: found 0 source and 1 binary packages D: opening db index /var/lib/rpm/Depends create mode=0x0 D: opening db index /var/lib/rpm/Providename rdonly mode=0x0 D: YES A oaf = 0.6.2-100_helix_2001020908 B oaf >= 0.2.4 D: Requires: oaf >= 0.2.4 YES (db provides) D: opening db index /var/lib/rpm/Basenames rdonly mode=0x0 D: Requires: /bin/sh YES (db files) D: NO A rpmlib(PayloadFilesHavePrefix) <= 4.0-1 B rpmlib(VersionedDependencies) = 3.0.3-1 D: NO A rpmlib(PayloadFilesHavePrefix) <= 4.0-1 B rpmlib(CompressedFileNames) = 3.0.4-1 D: NO A rpmlib(PayloadFilesHavePrefix) <= 4.0-1 B rpmlib(PayloadIsBzip2) = 3.0.5-1 D: YES A rpmlib(PayloadFilesHavePrefix) <= 4.0-1 B rpmlib(PayloadFilesHavePrefix) = 4.0-1 D: Requires: rpmlib(PayloadFilesHavePrefix) <= 4.0-1 YES (rpmlib provides) D: Requires: ld-linux.so.2 YES (db provides) D: Requires: libIIOP.so.0 YES (db provides) D: Requires: libORBit.so.0 YES (db provides) D: Requires: libORBitCosNaming.so.0 YES (db provides) D: Requires: libORBitutil.so.0 YES (db provides) D: Requires: libc.so.6 YES (db provides) D: Requires: libdl.so.2 YES (db provides) D: YES A libgconf-1.so.1 B libgconf-1.so.1 D: Requires: libgconf-1.so.1 YES (added provide) D: Requires: libglib-1.2.so.0 YES (db provides) D: Requires: libgmodule-1.2.so.0 YES (db provides) D: Requires: libm.so.6 YES (db provides) D: Requires: libnsl.so.1 YES (db provides) D: Requires: liboaf.so.0 YES (db provides) D: Requires: libpopt.so.0 YES (db provides) D: Requires: libxml.so.1 YES (db provides) D: Requires: libz.so.1 YES (dbprovides) D: Requires: libc.so.6(GLIBC_2.0) YES (db provides) D: Requires: libc.so.6(GLIBC_2.1) YES (db provides) D: Requires: libc.so.6(GLIBC_2.1.3) YES (db provides) D: NO A rpmlib(CompressedFileNames) <= 3.0.4-1 B rpmlib(VersionedDependencies) = 3.0.3-1 D: YES A rpmlib(CompressedFileNames) <= 3.0.4-1 B rpmlib(CompressedFileNames) = 3.0.4-1 D: Requires: rpmlib(CompressedFileNames) <= 3.0.4-1 YES (rpmlib provides) D: opening db index /var/lib/rpm/Conflictname rdonly mode=0x0 D: opening db index /var/lib/rpm/Requirename rdonly mode=0x0 D: YES A GConf = 0.12-100_helix_2001021007 B GConf = 0.12 D: Requires: GConf = 0.12 YES (added provide) D: Requires: R libgconf-1.so.1 YES (cached) D: Requires: R libgconf-1.so.1 YES (cached) D: YES A libgconf-gtk-1.so.1 B libgconf-gtk-1.so.1 D: Requires: libgconf-gtk-1.so.1 YES (added provide) D: Requires: R libgconf-gtk-1.so.1 YES (cached) D: Requires: R GConf = 0.12 YES (cached) D: installing binary packages D: getting list of mounted filesystems D: opening db index /var/lib/rpm/Group rdonly mode=0x0 D: opening db index /var/lib/rpm/Triggername rdonly mode=0x0 Segmentation fault (core dumped) Expected Results: I think you can figure this one out :) I installed the Nautlius rpm-4.0.2-0.34.i386.rpm with the build and python packages as well. http://www.eazel.com After I noticed that helix-update gave me core dumps I deinstalled this rpm and reinstalled the standard RH 7.0 rpm. Afterwards I did a --rebuilddb It says core dumped but no core is actually dumped. I have ulimit -c 10000
Ohh btw, after I had installed the original rpm from RedHat 7.0 I could install a few packages. But today everything seems impossible to install...
Dunno what Nautilus is shipping/using in rpm-4.0.2. There's a known problem with gcc mis-compiling rpm when -mcpu=i686 is used (just a guess at the cause). rpm depends on reliable data from the database, many segfaults in rpm are caused by unreliable data in the database. Specifically, after *any* segfault in rpm, you should do "rpm --rebuilddb", as further segfaults can be caused by unreliable data from the original segfault. What do you mean "impossible to install" ?
Nothing att all can be upgraded or installed as of now. Neither -i or -U works. I've tried to do --rebuilddb more than once now... /Richard
Ok, this is strange... I can do a -e and remove a package without a problem. Afterwards I do a --rebuilddb and still I can *not* do a -i, -U or -F but.... W8, I found something strange! When I try to do a rpm -Uvh *.rpm on the rpm packages provided each night by the Evolution-team I can't install, upgrade or freshen them. Not even when I try a single package. It *always* stops at: D: opening db index /var/lib/rpm/Group rdonly mode=0x0 D: opening db index /var/lib/rpm/Triggername rdonly mode=0x0 Segmentation fault BUT when I downloaded two other rpms it works to do a rpm -i. ftp://ftp.enst.fr/pub/unix/a2ps/a2ps-4.13-3.i386.rpm http://download.sourceforge.net/jedit/jedit-3.0.2-1.noarch.rpm So the above rpms work but the evolution nightly builds below do not work. ftp://ftp.ximian.com/pub/ximian/evolution-snapshots/distributions/RedHat-7/*.rpm This is so strange.. Can it really be the -mcpu=i686 bug? /Richard
ok, it's comfirmed. I can use all rpms made by RedHat, and others. But as soon as I try to install/upgrade/freshen rpms from HelixCode/Ximian I get segfaults. /Richard
I tried to replace the rpm, rpm-build, rpm-devel and rpm-python packages that comes with 7.0 with the ones that come with "fisher". That didn't work, I got a segfault, as I get with HelixCode/Evolution packages.. /Richard
Hmmm, I *think* that this is a duplicate of Bug 24692 ? /Richard
Okay, rpm finally gave me a core to work with. But I'm afraid that the backtrace didn't give much. [root@/]# gdb -c core /bin/rpm GNU gdb 5.0 Core was generated by `rpm -Uvh --test bonobo-0.36-100_helix_2001021008.i386.rpm bonobo-devel-0.36-100'. Program terminated with signal 11, Segmentation fault. #0 0x806e4d4 in ?? () (gdb) bt #0 0x806e4d4 in ?? () #1 0x8070371 in ?? () #2 0x8067974 in ?? () #3 0x804b29a in ?? () #4 0x8101801 in ?? () (gdb) Any ideas? I'm running out of option now... /Richard
Reinstall using rpm2cpio. As root do cd /var/tmp rpm2cpio rpm-4.0.1*.rpm | cpio -dim cp bin/rpm /bin and then reinstall rpm-4.0.1 packages using the executable that you installed manually. I'd suggest using rpm-4.0.1-0.7x from ftp://ftp.rpm.org/pub/rpm/test-4.0.1 or a recent Red Hat compiled rpm-4.0.2 from Raw Hide, that works for me. Please reopen if you still have problems.