[Date Prev][Date Next]
[Chronological]
[Thread]
[Top]
Re: slapd crash (ITS#38)
On Thu, Jan 14, 1999 at 08:39:54PM +0000, Kurt Zeilenga wrote:
> via AnonCVS. Would appreciate any testing you might be able to
> accomplish.
This does not help:
(gdb) r -d 769
Starting program: /home/hwr/src/openldap-rel-eng/ldap/servers/slapd/slapd -d
769
slapd 1.1.3-Release (Fri Jan 15 12:29:47 MET 1999)
hwr@shanagary:/home/hwr/src/openldap-rel-eng/ldap/servers/slapd
slapd starting
conn=0 fd=8 connection from localhost (127.0.0.1) accepted.
do_bind
conn=0 op=0 BIND dn="cn=Heiko Rupp,o=Xlink GmbH,c=DE" method=128
do_bind: version 2 dn (cn=Heiko Rupp,o=Xlink GmbH,c=DE) method 128
dn2entry_r: dn: cn=Heiko Rupp,o=Xlink GmbH,c=DE
=> dn2id( "cn=Heiko Rupp,o=Xlink GmbH,c=DE" )
=> ldbm_cache_open( "/var/db/ldap/dn2id.dbb", 514, 600 )
<= ldbm_cache_open (opened 0)
<= dn2id NOID
dn2entry_r: dn: o=Xlink GmbH,c=DE
=> dn2id( "o=Xlink GmbH,c=DE" )
=> ldbm_cache_open( "/var/db/ldap/dn2id.dbb", 514, 600 )
<= ldbm_cache_open (cache 0)
<= dn2id 1
=> id2entry_r( 1 )
=> ldbm_cache_open( "/var/db/ldap/id2entry.dbb", 514, 600 )
<= ldbm_cache_open (opened 1)
=> str2entry
<= str2entry 0x2e540
<= id2entry_r( 1 ) (disk)
====> cache_return_entry_r
send_ldap_result 0::
conn=0 op=0 RESULT err=0 tag=97 nentries=0
ber_get_next on fd 8 failed errno 35 (Resource temporarily unavailable)
*** got 98 of 120 so far
do_search
conn=0 op=1 SRCH base="" scope=1 filter="(!(objectclass=DSA))"
<= dereferencing dn
dn2entry_r: dn:
=> dn2id( "" )
=> ldbm_cache_open( "/var/db/ldap/dn2id.dbb", 514, 600 )
<= ldbm_cache_open (cache 0)
<= dn2id NOID
====> cache_return_entry_r
Program received signal SIGSEGV, Segmentation fault.
0x200ac in pthread_rdwr_runlock_np (rdwrp=0x14) at rdwr.c:49
49 if (rdwrp->readers_reading == 0) {
(gdb) bt
#0 0x200ac in pthread_rdwr_runlock_np (rdwrp=0x14) at rdwr.c:49
#1 0x70db in entry_rdwr_unlock (e=0x0, rw=0) at entry.c:271
#2 0x1399b in cache_return_entry_rw (cache=0x3a018, e=0x0, rw=0) at
cache.c:79
#3 0x139bc in cache_return_entry_r (cache=0x3a018, e=0x0) at cache.c:86
#4 0x1bb5e in derefDN (be=0x2a800, conn=0x2b4a0, op=0x2e400, dn=0x333d0 "")
at alias.c:207
#5 0x1a8b6 in ldbm_back_search (be=0x2a800, conn=0x2b4a0, op=0x2e400,
base=0x333d0 "", scope=1, deref=2, slimit=-1, tlimit=-1,
filter=0x39aa0,
filterstr=0x39c00 "(!(objectclass=DSA))", attrs=0x39be0, attrsonly=0)
at search.c:82
#6 0x41dd in do_search (conn=0x2b4a0, op=0x2e400) at search.c:148
#7 0x3776 in connection_operation (arg_v=0x37bb0) at connection.c:74
#8 0x201fb in pthread_create (tid=0x2e42c, attr=0xefbfd56c,
func=0x35d0 <connection_operation>, arg=0x37bb0) at thread.c:534
#9 0x3c94 in connection_activity (conn=0x2b4a0) at connection.c:225
#10 0x332b in slapd_daemon (port=0x185) at daemon.c:365
#11 0x201fb in pthread_create (tid=0x27ff4, attr=0x0,
func=0x238c <slapd_daemon>, arg=0x185) at thread.c:534
#12 0x1d5f in main (argc=3, argv=0xefbfd798) at main.c:139
(gdb) up
#1 0x70db in entry_rdwr_unlock (e=0x0, rw=0) at entry.c:271
271 return pthread_rdwr_runlock_np(&e->e_rdwr);
I did a configure --without-threads, as NetBSD mit-pthreads from
ftp://ftp.netbsd.org/pub/NetBSD/packages/pkgsrc/devel/mit-pthreads/README.html
are not liked by OpenLdap's configure:
setenv CPPFLAGS "-I/usr/pkg/include -I/usr/pkg/pthreads/include
-D_MIT_POSIX_THREADS"
setenv LDFLAGS "-L/usr/pkg/lib -L/usr/pkg/pthreads/lib"
checking POSIX thread version... final
checking for pthread_create in -lpthread... yes
checking for sched_yield... yes
checking for pthread_yield... yes
checking for pthread_kill... yes
checking for pthread_detach... yes
checking for pthread_attr_create... no
checking for pthread_attr_init... yes
checking for pthread_attr_destroy... yes
checking for pthread_attr_delete... no
checking for pthread_attr_setdetachstate... yes
checking for pthread_attr_setdetach_np... no
checking for pthread_setconcurrency... no
checking for thr_setconcurrency... no
checking if pthread_create() works... no
configure: error: pthread.h and pthread_create are not compatible
(But If I understand correctly, slapd does not necessarily require threads
anyway, else it should completely fail to compile).
Heiko