Re: (ITS#5701) connection.c asserting during test008

[hyc@symas.com]

ando@sys-net.it wrote:
> richton@nbcs.rutgers.edu wrote:
>
>> Rare connection.c assertion during test008s of RE24. Here are three different
>> examples:
>
>>    [6] monitor_subsys_conn_create(op = 0x103a359b0, rs = 0xffffffff76bff998, ndn
>> = (nil), e_parent = 0x1005c03d8, ep = 0xffffffff76bff230), line 500 in "conn.c"
>
>
>>    [6] monitor_subsys_rww_update(op = 0x102c3dd70, rs = 0xffffffff6dbff998, e =
>> 0x1005c1b98), line 187 in "rww.c"
>
>>    [7] monitor_subsys_rww_update(op = 0x100922640, rs = 0xffffffff78fff998, e =
>> 0x1005c1b98), line 185 in "rww.c"
>
>
> Sorry, I overlooked the fact that all those issues are occurring from
> within back-monitor.  In all cases, it's within code that loops through
> the connection array using connection_{first,next,close}().  Apparently,
> those calls are used in the right way.  Can you print more information
> about those connections?  E.g. c->c_struct_state, and *index from within
> connection_next()?  Or even better, the whole *c structure?

Looks like it may be related to #5586.
-- 
   -- Howard Chu
   CTO, Symas Corp.           http://www.symas.com
   Director, Highland Sun     http://highlandsun.com/hyc/
   Chief Architect, OpenLDAP  http://www.openldap.org/project/