Re: openldap23 w/ bdb43 entryCSN eq index prob

[Dusty Doris <openldap@mail.doris.cc>]

On Mon, 27 Jun 2005, Dusty Doris wrote:

> > When I restart in debug mode, I get a segmentation fault.
> >
> > => bdb_list_candidates 0xa0
> > => bdb_filter_candidates
> >         GE
> > => bdb_inequality_candidates (entryCSN)
> > => key_read
> > Segmentation fault
> >
>
> This actually gives me a core dump as well, if I wait a couple seconds.
> As soon as I figure out how to build a slapd binary with debugging symbols
> still intact, I can run gdb on that core file and post it.
>
> Anyway, this appears to be related to the syncprov overlay.  When I
> initialize the syncprov overlay in slapd.conf, I see this issue.  When I
> do not, then this doesn't happen.
>

Here is the core with gdb.  Is this useful?

(gdb) bt 20
#0  bdb_idl_fetch_key (be=0x8251300, db=0x824a000, tid=0x0,
key=0xbfb3e100,
    ids=0x837d000, saved_cursor=0xbfb3e180, get_flag=165) at idl.c:419
#1  0x283d9c55 in bdb_key_read (be=0x8251300, db=0x824a000, txn=0x0,
    k=0x827d004, ids=0x837d000, saved_cursor=0xbfb3e180, get_flag=165)
    at key.c:50
#2  0x283da8a4 in inequality_candidates (op=0xbfbfe710, ava=0xbfbfe610,
    ids=0x83fd000, tmp=0x837d000, gtorlt=165) at filterindex.c:1047
#3  0x283dac1c in bdb_filter_candidates (op=0xbfbfe710, f=0xbfbfe630,
    ids=0x83fd000, tmp=0x837d000, stack=0x847d000) at filterindex.c:168
#4  0x283dc187 in list_candidates (op=0xbfbfe710, flist=0xbfb3e4d0,
ftype=160,
    ids=0xbfb7e5b0, tmp=0x837d000, save=0x83fd000) at filterindex.c:506
#5  0x283dae38 in bdb_filter_candidates (op=0xbfbfe710, f=0xbfb3e500,
    ids=0xbfb7e5b0, tmp=0x837d000, stack=0x83fd000) at filterindex.c:189
#6  0x283d7630 in bdb_search (op=0xbfbfe710, rs=0xbfbfe6d0) at
search.c:1107
#7  0x080ac37a in glue_init ()
#8  0x080af494 in glue_init ()
#9  0x080a3274 in syncrepl_config ()
#10 0x0806bda2 in backend_startup_one ()
#11 0x0806bfa2 in backend_startup ()
#12 0x0805438b in main ()