[tao-bugs] Odp: Re: tao imr locator seg fault

Thu Jun 28 14:47:14 CDT 2018

 TAO VERSION: 2.4.6
 ACE VERSION: 6.4.6

HOST MACHINE and OPERATING SYSTEM:
Intel, cpu 16 cores, mem 128GB, Linux RHEL 7.3

COMPILER NAME AND VERSION (AND PATCHLEVEL):
gcc 8.1

THE $ACE_ROOT/ace/config.h FILE:
ace/config-linux.h

THE $ACE_ROOT/include/makeinclude/platform_macros.GNU FILE
include/makeinclude/platform_macros.GNU
Additional flags:
debug=0
ssl=1
zlib=1
bzip2=1
c++14=1
install_rpath=0
CPP_FLAGS += -DACE_ANY_OPS_USE_NAMESPACE

SYNOPSIS:
segmentation fault of "tao imr locator"

DESCRIPTION:

The following is an OCR-ed subset of previously sent png images:
gdb's backtrace:

1 Program received signal SIGSEGV, Segmentation fault.
2 0x00007f34abfd1257 in LiveEntry::validate_ping(bool&, ACE_Time_Value&)
3 from /tao/6.4.6/lib/libTAO_1mR_Locator.so.2.4.6
4 (gdb) where
5 #0 0x00007f34abfd1257 in LiveEntry::validate_ping(bool&, ACE_Time_Value&) () from /tao/6.4.6/lib/libTAO_1mR_Locator.so.2.4.6
6 #1 0x00007f34abfd1ao1 in LiveCheck::handle_timeout(ACE_Time_Value const&, void const*) () from /tao/6.4.6/lib/libTAO_1mR_Locator.so.2.4.6
7 #2 0x00007f34a92eaa95 in ACE_Ęvent_Handler_Handle_Timeout_Upcall::timeout(ACE;Abstract_Timer_Queue<ACE_Ęvent_Handler*>&, ACE_Ęvent_Handler*, void const*,
int, ACE_Time_Value const&) () from /tao/6.4.6/lib/libACE.so.6.4.6
8 #3 0x00007f34a96c615c in ACE_Iimer_Queue_T<ACE_Ęvent_Handler*, ACE_Ęvent_Handler_Handle_Timeout_Upcall, ACE_Recursive_Thread;Mutex,
ACE_System;Time_Policy>::expire_single(ACE_Command;Base&) () from /tao/6.4.6/lib/libTAO.so.2.4.6
9 #4 0x00007f34a92e54fd in ACE_Dev_Poll_Reactor::dispatch_timer_handler(ACE_Dev_Poll_Reactor::Token_Guard&)
10 from /tao/6.4.6/lib/libACE.so.6.4.6
11 #5 0x00007f34a92e6711 in ACE_Dev_Poll_Reactor::dispatch(ACE_Dev_Poll_Reactor::Token_Guard&) () from /tao/6.4.6/lib/libACE.so.6.4.6
12 #6 0x00007f34a92e7d2e in ACE_Dev_Poll_Reactor::handle_events(ACE_Time_Value*) () from /tao/6.4.6/lib/libACE.so.6.4.6
13 #7 0x00007f34a969168d in TAO_ORB_Core::run(ACE_Time_Value*, int) () from /tao/6.4.6/lib/libTAO.so.2.4.6
14 #8 0x00007f34abfca580 in ImR_Locator_i::run() () from /tao/6.4.6/lib/libTAO_1mR_Locator.so.2.4.6
15 #9 0x0000000000401d16 in run_standalone(options&)
16 #10 0x0000000000401b3f in main
17 (gdb)
18

It looks like this concerns the tao pinging mechanism. With the pinging mechanism (-v 0) I do not observe the segmentation fault.
The following shows all parameters when starting up the services:

$ ps aux lgrep tao

dev 7633 0.0 0.0 59128 7100 ? S Jun24 0:00 /tao/6.1.8/bin/tao_cosnaming -ORBListenEndPoints iiop://l.2@host1 -ORBDefaultInitRef corbaloc:iiop:1.2 at host1:59695 -ORBUseIMR 1 -ORBLogFile /tao/6.1.8/log/Naming_5ervice.log -u /tao/6.1.8/Name5ervice -o /tao/6.1.8/name_service.ior
dev 23112 0.4 0.3 983116 922884 ? Ss Jun24 19:56 /tao/6.1.8/bin/tao_imr_locator -ORBListenEndpoints iiop://l.2@host1:59695 -x /tao/6.1.8/ImpRep/impl_repo.xm1 -ORBSvcConf /svc.conf.poll -ORBVerboseLogging 1
dev 23117 0.0 0.0 64012 7504 ? Ss Jun24 0:11 /tao/6.1.8/bin/tao_imr_activator -l -ORBDefaultInitRef corbaloc:iiop:1.2 at host1:59695 -ORBSvcConf /svc.conf.poll -ORBVerboseLogging 1

Regards,
Tomasz Mikolajczyk

Dnia Czwartek, 28 Czerwca 2018 08:05 Johnny Willemsen <jwillemsen at remedy.nl> napisał(a) 
> Hi,
> 
> Please us the problem-report-form that is in the root of the package and
> provide the output as txt, looks the png images are dropped by the list
> server but these also make it hard to do so some quick searches.  It
> would help when you can extend/add a unit test under
> TAO/orbsvcs/tests/ImplRepo to reproduce your problem.
> 
> Best regards,
> 
> Johnny Willemsen
> Remedy IT
> Postbus 81 | 6930 AB Westervoort | The Netherlands
> http://www.remedy.nl
> 
> On 06/27/2018 11:37 PM, Tomasz Mikolajczyk wrote:
> > Hi,
> >
> > I encountered TAO imr locator seg fault. PFA all the possible information collected. Unfortunately due to technical reasons I'm not able to send them as a single text file. Tao processes are running on localhost. processes just talk to themselves executing simple methods defined in IDL files.
> >
> > Probably the most interested files at the beginning will be:
> > tao_seg_fault01.png - gdb's backtrace 
> > tao_seg_fault21.png  - shows options passed when starting the tao processes
> >
> > I was comparing the changes between ACE+TAO 6.4.7 version and ACE+TAO  6.5 and I do not see any potential fix for this issue.
> >
> > Should you need further details please let me know.
> >
> > Regards,
> > Tomasz Mikolajczyk
> >
> >
> > _______________________________________________
> > tao-bugs mailing list
> > tao-bugs at list.isis.vanderbilt.edu
> > http://list.isis.vanderbilt.edu/cgi-bin/mailman/listinfo/tao-bugs