Red Hat Bugzilla – Bug 47166
core dump with rpm -Uvh
Last modified: 2007-04-18 12:34:25 EDT
From Bugzilla Helper:
User-Agent: Mozilla/4.76 [en] (X11; U; Linux 2.2.18pre17 i686)
Description of problem:
If I Install or Upgrade the package ntp (downloaded it from 3 different
locations and every time the same), I get a core dump..., I already
installed new binary rpm's, doesn';t help, also did a rpm --rebuilddb
Steps to Reproduce:
1. rpm -Uvh ntp*rpm
Actual Results: a core file
package NOT installed
Expected Results: the package to be installed
A pointer (i.e. URL, bugzilla attachments won't work) to a copy of your database
tar czvf /tmp/rpmdb.tar.gz rpm
also interessting, when I delete the entire db, and do a rpm --initdb, I still
got the problem...
Hmmm, I had assumed that your problem was due to getting
corrupted data from the database, that's not the case AFAICT.
What version of rpm are you using? I've just put out a final
test release for rpm-4.0.3-0.57 at
Can rpm-4.0.3-0.57 install your ntp package?
Otherwise please give me a pointer to the ntp package that
has the problem, and I'll take a look. Thanks.
Please reopen if a "rpm --rebuild" with rpm-4.0.3-0.57 does
not fix your problem.
I appeared to have fixed the problem
It doesnt matter what package I use of rpm, whatever kind of version,
and even whatever package I'm installing, I get it continuously, but why?
well, I managed to trace it:
<tfm> (gdb) r
<tfm> Starting program: /tmp/recover/build/bin/rpm -Uvh
<tfm> Preparing... ###########################################
<tfm> Program received signal SIGSEGV, Segmentation fault.
<tfm> 0x4055309a in ?? ()
<tfm> #21 0x08129876 in __getpwnam_r (name=0x816afb9 "root", resbuf=0x81a4684,
<tfm> big.gif", buflen=1024, result=0xbffff558) at ../nss/getXXbyYY_r.c:200
<tfm> #22 0x0812944b in getpwnam (name=0x816afb9 "root") at
<tfm> #23 0x0805d073 in installBinaryPackage (ts=0x81b9be8, fd=0x81ad938,
<tfm> h=0x81c4fa8, pkgKey=0xbffffc4b, actions=0x81c3898, sharedList=0x0)
<tfm> at install.c:1044
<tfm> #24 0x08070be9 in rpmRunTransactions (ts=0x81b9be8,
<tfm> notify=0x8066b3c <showProgress>, notifyData=0x36, okProbs=0x0,
<tfm> newProbs=0xbffff928, transFlags=0, ignoreSet=116) at
<tfm> #25 0x080678eb in rpmInstall (rootdir=0x8166bac "/", fileArgv=0x81a76d8,
<tfm> transFlags=0, interfaceFlags=38, probFilter=116, relocations=0x0)
<tfm> at rpminstall.c:460
<tfm> #26 0x0804b1b7 in main (argc=5, argv=0xbffffb24) at rpm.c:1146
<tfm> #27 0x08101a32 in __libc_start_main (main=0x8049840 <main>, argc=5,
<tfm> ubp_av=0xbffffb24, init=0x80480b4 <_init>, fini=0x8163be0 <_fini>,
<tfm> rtld_fini=0, stack_end=0xbffffb1c) at
so what was the problems:
in /etc/nsswitch.conf, I had:
passwd: ldap files nisplus nis
I replaced it with:
passwd: files nisplus nis ldap
and now it works perfectly
Still I find that it is WRONG to core dump on this, this is a bug and it should
btw, my passwd file is completely blank except for the root account
Agreed no core dump. Off to LDAP for resolution (sorry Nalin :-)
Actually, I know this one already (thank you Nalin!).
Make sure that nscd is running.
hmmm, it wasn't running, now it is
a) this should be mentioned ;)
b) core dumping for this one ?
real problem I had was that rpm really didn;t give me ANY info, even not in a
trivial tool like strace, I had to compille rpm on my own, so I had all the
debugging info, and then I could do a decent gdb on it, mja, anyway, my
m,ailserver is back rpm compatible and that is what counts
tx for the help btw ;)