[Vm-dev] Freeze after Morph Activity

[Dan Norton]

On 01/29/2017 12:11 AM, Ben Coman wrote:
>   
> Highly speculative idea... from a shell could you log your system time
> to see if anything there correlates.
Hi Ben,

This is what I did:

# tail -f /var/log/messages > /home/dan/systime.log

Up to a freeze, this was written:

Jan 29 09:52:51 debian kernel: [   12.957276] IPv6: 
ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
Jan 29 09:52:55 debian org.a11y.Bus[1050]: Activating service 
name='org.a11y.atspi.Registry'
Jan 29 09:52:55 debian org.a11y.Bus[1050]: Successfully activated 
service 'org.a11y.atspi.Registry'
Jan 29 09:52:55 debian org.a11y.atspi.Registry[1056]: SpiRegistry daemon 
is running with well-known name - org.a11y.atspi.Registry
Jan 29 09:53:08 debian org.a11y.Bus[1127]: Activating service 
name='org.a11y.atspi.Registry'
Jan 29 09:53:08 debian org.a11y.Bus[1127]: Successfully activated 
service 'org.a11y.atspi.Registry'
Jan 29 09:53:08 debian org.a11y.atspi.Registry[1182]: SpiRegistry daemon 
is running with well-known name - org.a11y.atspi.Registry
Jan 29 09:57:50 debian rsyslogd: [origin software="rsyslogd" 
swVersion="8.4.2" x-pid="859" x-info="http://www.rsyslog.com"] rsyslogd 
was HUPed
Jan 29 10:07:44 debian rsyslogd0: action 'action 17' resumed (module 
'builtin:ompipe') [try http://www.rsyslog.com/e/0 ]
Jan 29 10:07:44 debian rsyslogd-2359: action 'action 17' resumed (module 
'builtin:ompipe') [try http://www.rsyslog.com/e/2359 ]

  - Dan
> cheers -ben
>
> On Sun, Jan 29, 2017 at 11:40 AM, Dan Norton <dnorton at mindspring.com> wrote:
>>
>> Hi Levente,
>>
>> I directed stdout to a file and prior to the freeze, did:
>>
>> pkill -USR1 -n -x squeak
>>
>> a few times and noticed "stack overflow" mentioned twice. After the freeze,
>> pkill sent nothing else to stdout. Maybe something in stderr? I'll try
>> again and concatenate stderr to the end of stdout. Meanwhile, the file with
>> the stack overflows is attached.
>>
>> Thanks for the help.
>>
>>   - Dan
>>
>> On Sat, 28 Jan 2017, Levente Uzonyi wrote:
>>
>>
>>> Hi Dan,
>>> You can send the USR1 signal to the VM process to make it write some debug
>>> information to the console. With this information you can easily tell
>>> what's happening. I suspect it's stuck in a long GC.
>>> Levente
>>> On Fri, 27 Jan 2017, Dan Norton wrote:
>>>> On Debian8, CogSpur64 5.0-201612221637, Cuis 5.0 3043...
>>>>
>>>> after lots of user interaction, views opening and closing, and
>>>> animation, the image becomes unresponsive. Cmd+. does nothing and the
>>>> clock in the Cuis taskbar no longer updates. The length of time to
>>>> produce this varies from 5 to 20 minutes.
>>>>
>>>> There seems to be no dump and no log. 'Smalltalk garbageCollectMost'
>>>> reports 25550736 to 30517872 over 34 samples.
>>>>
>>>> Sorry to be so vague. This has occurred with several images - sometimes
>>>> scrolling through a senders list, or stepping through a debugger, but
>>>> repeatably with one of my images. I can supply this image, warmed up so
>>>> that it might not take so long to reproduce the problem if you would
>>>> like. Or give me a hint as to how to narrow down the problem
>>