Re: [Fwd: tests/rr-server core dumps multi-processor]

oops spoke too soon!

Still fails, included below core dumps from rr-server and rr-client 
(when rr-server stays up)

#0  0xfeeb3344 in strlen () from /usr/lib/libc.so.1
#1  0xfef03098 in _doprnt () from /usr/lib/libc.so.1
#2  0xfef0520c in vsnprintf () from /usr/lib/libc.so.1
#3  0xff233b14 in g_logv (log_domain=0xff30cc50 "GLib-GObject",
     log_level=G_LOG_LEVEL_WARNING,
     format=0xff30e8b8 "invalid uninstantiatable type `%s' in cast to 
`%s'",
     args1=0xfe909b08) at gmessages.c:436
#4  0xff233e90 in g_log (log_domain=0xff30cc50 "GLib-GObject",
     log_level=G_LOG_LEVEL_WARNING,
     format=0xff30e8b8 "invalid uninstantiatable type `%s' in cast to `%s'")
     at gmessages.c:527
#5  0xff300514 in g_type_check_instance_cast (type_instance=0x40228,
     iface_type=80) at gtype.c:2788
#6  0x00011df4 in frame_available (channel=0x457b8, frame=0x40228,
     error=0xfe909c0c) at echo.c:141
#7  0xff359978 in do_frame_available (data=0x457b8, user_data=0x40228)
     at rr-channel.c:367
#8  0xff36428c in work_proxy (data=0x44c20) at rr-workpool.c:143
#9  0xff246c6c in g_thread_create_proxy (data=0x0) at gthread.c:551

Occasionally get core from client (and server stays running)
#0  0xfefd9794 in __sigprocmask () from /usr/lib/libthread.so.1
#1  0xfefce9a8 in _resetsig () from /usr/lib/libthread.so.1
#2  0xfefce148 in _sigon () from /usr/lib/libthread.so.1
#3  0xfefd1188 in _thrp_kill () from /usr/lib/libthread.so.1
#4  0xfeecb760 in raise () from /usr/lib/libc.so.1
#5  0xfeeb5a84 in abort () from /usr/lib/libc.so.1
#6  0xff233de8 in g_logv (log_domain=0x0, log_level=G_LOG_LEVEL_ERROR,
     format=0xfe6036d8 "rr_null_echo_trip: Channel closed.\n", 
args1=0x3e6a0)
     at gmessages.c:508
#7  0xff233e90 in g_log (log_domain=0x0, log_level=G_LOG_LEVEL_ERROR,
     format=0x12578 "rr_null_echo_trip: %s\n") at gmessages.c:527
#8  0x000112b4 in do_tests (connection=0x44c48) at rr-client.c:37
#9  0x00011390 in my_thread (data=0x2b600) at rr-client.c:55
#10 0xff246c6c in g_thread_create_proxy (data=0x0) at gthread.c:551

Paul Caffrey wrote:
> Hi
> 
> Okay it appears to work now!
> 
> What I found was that glib was built on a single cpu machine
> 
> By building it on a multi-cpu machine it then runs okay on a multi-cpu 
> machine
> 
> Regards
> 
> Paul
> 
> 
> ------------------------------------------------------------------------
> 
> Subject:
> tests/rr-server core dumps multi-processor
> From:
> Paul Caffrey <[email protected]>
> Date:
> Tue, 18 Feb 2003 12:01:13 +0000
> To:
> [email protected]
> 
> 
> Hi
> 
> What I've discovered is that it works ok (so far) on a single cpu box
> 
> However on a 4 cpu machine it hangs within about 5 minutes
> 
> Paul
> 
> 
> ------------------------------------------------------------------------
> 
> Subject:
> Re: [Roadrunner] tests/rr-server core dumps
> From:
> Paul Caffrey <[email protected]>
> Date:
> Tue, 18 Feb 2003 10:03:10 +0000
> To:
> Jonas Borgström <[email protected]>
> 
> 
> Hi Jonas
> 
> Jonas Borgström wrote:
> 
>> On Mon, 2003-02-17 at 18:26, Paul Caffrey wrote:
>>
>>> Hi
>>>
>>> We've tried to use beepcore but found it very buggy.
>>>
>>> We next tried RoadRunner hoping it would be more stable
>>>
>>
>>
>> This is strange, it should be :)
>>
>>
>>> However we find the sample tests core dump and hang out of the box!
>>> (rr-server can core dump, while rr-client can hang)
>>>
>>
>> Can you provide some more information to make the debugging easier:
>>
>> Which OS and OS-version are you using?
> 
> Solaris 5.8 Generic_108528-15 sun4u sparc SUNW,Ultra-4
> 
>> Which glib2 version?
> 
> Name: GLib
> Description: C Utility Library
> Version: 2.2.1
> 
>> Which libxml2 version?
> 
> Name: libXML
> Version: 2.5.0
> 
>> Which RoadRunner version?
> 
> Name: librr
> Description: a BEEP implementation
> Requires: glib-2.0 gobject-2.0 gthread-2.0 libxml-2.0
> Version: 0.9.1
> 
>>
>> / Jonas
>>
>> PS
>> Can you please join the mailing list so I don't have to manually approve
>> your postings.
> 
> Done!
> 
>> DS
> 
> 
>