RP/TUX 6.5, TUX 6.51 - Solaris - GWTDOMAIN dumps a core (Doc ID 773807.1)

Last updated on NOVEMBER 04, 2016

Applies to:

Oracle Tuxedo
Information in this document applies to any platform.
Information in this document applies to any platform
***Checked for relevance on 25-07-2011***

Goal

DESCRIPTION:
Timing might cause GWTDOMAIN to dump a core.

CONFIGURATION:
export GW_VALIDATE_HOST=YES
tmboot -A -s DMADM -s GWADM
truss -u \* -u !libc,libc_psr -a -o GWT.out GWTDOMAIN -g 10 -i 31 -u frsol03 -U /tmp/ULOG -m 0 -A
    or
I have also obtained the same core letting the configuration up overnight

The customer is reproducing this each time on his low-end platform with fast network. Slowing the network works around
this.

Stack trace looks like
program terminated by signal BUS (invalid address alignment)
    or
program terminated by signal SEGV (no mapping at the fault address)
(/home0/Forte6/opt/SUNWspro/bin/../WS6U1/bin/sparcv9/dbx) where
=>[1] gw_nw_getconnect(0x0, 0x5, 0x31708, 0x31700, 0x8, 0x22070), at 0xff2e5f04
  [2] gw_nw_acall(0x22070, 0x8, 0x2e728, 0x2e510, 0x31700, 0xff380720), at 0xff2e612c
  [3] gw_run_svr(0xff380728, 0x22070, 0xff37e874, 0x2fed0, 0x2e510, 0x0), at 0xff3324a4
  [4] _tmrunserver(0x22070, 0x2fed0, 0x0, 0xff290644, 0xff28e000, 0x281e8), at 0x118e0
  [5] _tmstartserver(0xff28e000, 0x0, 0x21f94, 0x22140, 0x0, 0x0), at 0xff236688

truss ends with
      -> libgpnet:_gp_nwhosts2n(0xffbef470, 0xffbef470, 0xffbef48c, 0xff30b30c)
        -> libnsl:gethostbyname(0xffbef470, 0x0, 0x0, 0x0)
        -> libnsl:gethostbyname_r(0xffbef470, 0x719cc, 0x719e0, 0x920)
door_info(5, 0xFFBED1D0)                        = 0
door_call(5, 0xFFBED1B8)                        = 0
        <- libnsl:gethostbyname() = 0x719cc
time()                                          = 1062608632
        -> libnsl:inet_ntoa(0xffbef3fc, 0x0, 0x3f561ef8, 0x920)
        <- libnsl:inet_ntoa() = 0xff0a3b54
      <- libgpnet:_gp_nwhosts2n() = 0x6ef30
    Incurred fault #5, FLTACCESS  %pc = 0xFF2E5F04
      siginfo: SIGBUS BUS_ADRALN addr=0x6572697B
    Received signal #10, SIGBUS [default]
      siginfo: SIGBUS BUS_ADRALN addr=0x6572697B
        *** process killed ***

Solution

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms