On Thu, Feb 23, 2012 at 7:07 AM, Harezga, Nick <nhare...@mitre.org> wrote:
> Would this option allow for me to checkpoint an IRC daemon when the clients 
> connected to it aren't also being checkpointed? That was my interpretation of 
> the description contained in the --help flag in the configure script. If that 
> is incorrect, could you please give a proper explanation? If this isn't the 
> correct method to handle network connections where both sides aren't 
> checkpointed, is there a way to accomplish this?

That flag will _not_ help you in checkpointing the server with clients
connected to it. That flag just allows DMTCP to wait until all the
external sockets have disconnected, so I don't think this is the
desired policy. In your situation, we need to ignore the external
sockets and continue to checkpoint.  There might be a simple way, but
I will have to test it. I will test it locally and will write you
back.

Can you tell me how to reproduce the problem. Which IRC server should
I be using and which client :). Also, I assume that X11 is not
involved in all this, am I correct?

Thanks,
Kapil

>
> -----Original Message-----
> From: Kapil Arya [mailto:ka...@ccs.neu.edu]
> Sent: Thursday, February 23, 2012 12:03 AM
> To: Gene Cooperman
> Cc: Harezga, Nick; dmtcp-forum@lists.sourceforge.net
> Subject: Re: [Dmtcp-forum] Compile error caused by configure flag
>
> Hi Nick,
>
> Thanks for reporting the bug. The configure flag
> "--enable-external-socket-handling" pertains to an older piece of
> code. This option hasn't been exercised lately and is broken due to
> benign negligence. It has been on our todo list to replace this code
> with a newer design that uses DMTCP modules but has been low priority
> until now. If this flag is desired, then we will be happy to raise the
> priority and provide you with the solution over the next few days.
>
> Is this something you would want to have in DMTCP rather sooner than
> later? If so, please let us know and we will find a workaround.
>
> Thanks,
> Kapil
>
> On Wed, Feb 22, 2012 at 12:11 PM, Gene Cooperman <g...@ccs.neu.edu> wrote:
>> Thanks for sending this report, Nick.  We'll get back to you soon.
>>
>> Best wishes,
>> - Gene
>>
>> On Wed, Feb 22, 2012 at 12:25:54PM +0000, Harezga, Nick wrote:
>>> Hi all,
>>>
>>> After talking to Professor Cooperman yesterday, he recommended that I 
>>> e-mail the list with the issue that I am having. For some background, I am 
>>> attempting to checkpoint an IRC server (Unreal), which introduces external 
>>> sockets. This works fine when I also checkpoint clients connected to the 
>>> server, but that isn't what I am trying to do.
>>>
>>> I tried to configure DMTCP using: "./configure CPPFLAGS="-DDEBUG" 
>>> CXXFLAGS="-O0 -g3" --enable-debug --enable-external-socket-handling"
>>>
>>> I wasn't able to successfully checkpoint when the client isn't being 
>>> checkpointed, so I figured I would try the external-socket-handling flag to 
>>> see if that did anything. The configure works fine, but after running make, 
>>> I receive the following error:
>>>
>>> g++ -DHAVE_CONFIG_H -I.   -DDEBUG -fPIC -g -O0 -MT connectionstate.o -MD 
>>> -MP -MF .deps/connectionstate.Tpo -c -o connectionstate.o 
>>> connectionstate.cpp
>>> In file included from constants.h:26,
>>>                  from connectionstate.cpp:22:
>>> config.h:8:1: warning: "DEBUG" redefined
>>> <command-line>: warning: this is the location of the previous definition
>>> In file included from connectionmanager.h:27,
>>>                  from connectionstate.h:27,
>>>                  from connectionstate.cpp:23:
>>> connection.h:157: error: 'TcpConnectionInfo' was not declared in this scope
>>> connection.h:157: error: template argument 1 is invalid
>>> In file included from connectionstate.cpp:23:
>>> connectionstate.h:51: error: 'TcpConnectionInfo' was not declared in this 
>>> scope
>>> connectionstate.h:51: error: template argument 1 is invalid
>>> connectionstate.cpp:151: error: 'TcpConnectionInfo' was not declared in 
>>> this scope
>>> connectionstate.cpp:151: error: template argument 1 is invalid
>>> make[3]: *** [connectionstate.o] Error 1
>>> make[3]: Leaving directory `/home/nharezga/dev/dmtcp/dmtcp-1.2.4/dmtcp/src'
>>> make[2]: *** [all] Error 2
>>> make[2]: Leaving directory `/home/nharezga/dev/dmtcp/dmtcp-1.2.4/dmtcp/src'
>>> make[1]: *** [all-recursive] Error 1
>>> make[1]: Leaving directory `/home/nharezga/dev/dmtcp/dmtcp-1.2.4/dmtcp'
>>> make: *** [dmtcp] Error 2
>>>
>>> I am using the latest SVN revision, 32 bit Ubuntu 10.04 in a virtual 
>>> machine, and kernel 2.6.32-38.
>>>
>>> Any ideas?
>>
>>> ------------------------------------------------------------------------------
>>> Virtualization & Cloud Management Using Capacity Planning
>>> Cloud computing makes use of virtualization - but cloud computing
>>> also focuses on allowing computing to be delivered as a service.
>>> http://www.accelacomm.com/jaw/sfnl/114/51521223/
>>
>>> _______________________________________________
>>> Dmtcp-forum mailing list
>>> Dmtcp-forum@lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/dmtcp-forum
>>
>>
>> ------------------------------------------------------------------------------
>> Virtualization & Cloud Management Using Capacity Planning
>> Cloud computing makes use of virtualization - but cloud computing
>> also focuses on allowing computing to be delivered as a service.
>> http://www.accelacomm.com/jaw/sfnl/114/51521223/
>> _______________________________________________
>> Dmtcp-forum mailing list
>> Dmtcp-forum@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/dmtcp-forum

------------------------------------------------------------------------------
Virtualization & Cloud Management Using Capacity Planning
Cloud computing makes use of virtualization - but cloud computing 
also focuses on allowing computing to be delivered as a service.
http://www.accelacomm.com/jaw/sfnl/114/51521223/
_______________________________________________
Dmtcp-forum mailing list
Dmtcp-forum@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dmtcp-forum

Reply via email to