Re: Bus Error - core dumped

["Bharat Tewari" <tewari@wipinfo.soft.net>]

hi rosdi
well seems you are running out of memory coz the malloc is failing, you are
not able to get continous chunk of memory, i had the same problem with some
other implementation of ldap, my suggestion first check if you can increase
the ulimit.
do ulimit -a and check on vmemory, if its set to some value change it like
this
ulimit -v unlimited (if you have super user access its going to work)
secondly if it still core dumps, then do the following increase the size of
chunk from 120 to some huge variable like 1024 or 2048 types, check in the
code where ch_calloc() is called in your ldap
i.e file ch_malloc.c and try oops just now saw that in frame #9 , the base
value is out of bounds
in shell_back_search()??????? anyone worked with shell_back_search(),
shouldnt this contain the
base value.
my 1-bit thoughts.
bye
bharat

----- Original Message -----
From: Rosdi Kasim <rosdi@snt.com.my>
To: <openldap-software@openldap.org>
Sent: Thursday, December 23, 1999 12:00 PM
Subject: Re: Bus Error - core dumped


> I'm still having problem with OpenLDAP (1.2.7 or 1.2.8 - I have tried
both)
> under Sun SPARC.
> I use gdb to debug the program, and here is the stack trace slapd dumped
to
> me...
>
> (gdb) bt
> #0  0xef5c6694 in t_splay () from /usr/lib/libc.so.1
> #1  0xef5c6504 in t_delete () from /usr/lib/libc.so.1
> #2  0xef5c61d8 in realfree () from /usr/lib/libc.so.1
> #3  0xef5c5d88 in _malloc_unlocked () from /usr/lib/libc.so.1
> #4  0xef5c5b38 in malloc () from /usr/lib/libc.so.1
> #5  0xef5baca8 in calloc () from /usr/lib/libc.so.1
> #6  0x216e8 in ch_calloc (nelem=1, size=120) at ch_malloc.c:56
> #7  0x1bb10 in str2entry (
>     s=0x5f100 "dn:
> uid=rosdi,ou=Network/Internet,ou=Staff,o=SNT,st=Sarawak,c=MY\
> ncn: Rosdi Kasim\ncreatetimestamp: 19990625082001Z\ncreatorsname:
> uid=admin,st=S
> arawak,c=MY\nfacsimiletelephonenumber: 03-2630408\ngivenname"...) at
> entry.c:63
> #8  0x297a8 in read_and_send_results (be=0x51108, conn=0x59c18,
op=0x52788,
>     fp=0x4d9c0, attrs=0x52c98, attrsonly=0) at result.c:57
> #9  0x29ab0 in shell_back_search (be=0x51108, conn=0x59c18, op=0x52788,
>     base=0x2c <Address 0x2c out of bounds>, scope=240640, deref=3,
size=100,
>     time=60, filter=0x58530,
>     filterstr=0x52848
> "(|(mail=ROSDI*)(|(cn=ROSDI*)(|(sn=ROSDI*)(givenname=ROSDI
> *))))", attrs=0x52c98, attrsonly=0) at search.c:67
> #10 0x19254 in do_search (conn=0x59c18, op=0x52788) at search.c:149
> #11 0x1870c in connection_operation (arg_v=0x58cd0) at connection.c:74
> (gdb)
>
>
> Being not a C or C++ programmer myself, the whole thing seems rubbish to
me,
> can anybody in here give me a clue as what should I do next?  I hate to
say
> this, but I have to be honest, "I'm desperate.."
>
>
> Rosdi.
>
>
> ----- Original Message -----
> From: "Bharat Tewari" <tewari@wipinfo.soft.net>
> To: "Rosdi Kasim" <rosdi@snt.com.my>; <openldap-software@openldap.org>
> Sent: Wednesday, December 22, 1999 11:24 AM
> Subject: Re: Bus Error - core dumped
>
>
> > bus error is caused mainly due to the misaligned data members i.e if int
> > starts at an odd memory location then accessing it as int would cause
bus
> > error, now i run ldap on sparc solaris 2.7 and i dont get any error such
> as
> > this so i am not really sure oops i work with 1.2.7 and not 1.2.8, see
you
> > can
> > do this, you can compile it with "-g" (i guess this is default) run
slapd
> > under dbx i.e
> > start slapd normally and then do
> > dbx - <process id of slapd> and then at the prompt type cont
> > now do ldapsearc from the client and give back the stack trace and the
> > problem will be fixed easily
> > bye
> > bharat
>
>
>