Re: A first look at diskio for HP-UX

[Didier C. <did...@fe...>]

Hello,

    I use ucd-snmp V4.1.1. I've compiled a master agent on SunOS5.7 that
support host MIB and MIB II.
Our validation department is making some load test on this particular
master agent.

We run the master agent on a SunOS5.7.
Then we run the following command on a remote linux :
while [ 1 = 1 ]; do snmpwalk [mastergent_machine_name] public; done

After 20 hours the master agent crash.
Here is the stack obtained by using dbx on core dump.

program terminated by signal SEGV (no mapping at the fault address)
(/opt/SUNWspro/bin/../WS5.0/bin/sparcv9/dbx) where
=>[1] __align_cpy_1(0xffbeccf0, 0x17bff8, 0x18, 0x4, 0x10, 0xffbeccc0),
at 0xff1e0758
  [2] getMibstat(0x4, 0xffbeccc0, 0x48, 0x2, 0x36d98, 0xffbecd08), at
0x57230
  [3] var_ipAddrEntry(0xffbed008, 0x13bdd0, 0x13bdc0, 0x0, 0xffbed524,
0xffbed51c), at 0x36fe0
  [4] search_subtree_vars(0x131130, 0x13bdd0, 0x13bdc0, 0xffbed293,
0x14, 0x131558), at 0x2915c
  [5] getStatPtr(0xffbed51c, 0x13a878, 0xffbed52f, 0xffbed524,
0xffbed522, 0x0), at 0x29a54
  [6] handle_var_list(0xffbed51c, 0x5, 0xffbed59c, 0xffbed4d0, 0x0,
0x12d270), at 0x283e8
  [7] handle_next_pass(0x13a7c0, 0x13a878, 0xffbed798, 0x33, 0xff1b3938,
0xff145590), at 0x2820c
  [8] handle_snmp_packet(0x1, 0x13a580, 0x4c109d3a, 0x13a680, 0x0,
0x13a7b5), at 0x27a20
  [9] _sess_read(0x12de60, 0xffbef918, 0x0, 0x0, 0x0, 0x0), at 0x7b350
  [10] snmp_sess_read(0x12de60, 0xffbef918, 0x1, 0x0, 0xff1b69cc,
0xffffffff), at 0x7b418
  [11] snmp_read(0xffbef918, 0xff1b69cc, 0xff1b69cc, 0xff1b69cc,
0xff1b69cc, 0x8), at 0x79e98
  [12] receive(0x0, 0x0, 0xffffffff, 0xfffffff8, 0x0, 0x116e49), at
0x269a8
  [13] main(0x5, 0xffbefca4, 0xffbefcbc, 0x108800, 0x0, 0x0), at 0x2660c

Have you got any idea ? Have you ever seen this kind of problem ?

Regards,

Didier Chabanol
FERMA