Dear Kapil,
I went through the paper and here is my understanding of it: It essentially
block all the processes till the checkpointing is complete which ensures
that there are no orphan messages in the saved state. Am I correct?
We are working on developing non-blocking or partial blocking checkpoint
metho <http://link.springer.com/chapter/10.1007%2F978-3-642-45249-9_9>ds; I
was looking for a way to implement our algorithms. I believe that for the
same, I will have to change the code of DMTCP in order to stop blocking.
Best Regards,
Pushpendra Singh
http://www.iiitd.edu.in/~pushpendra
On Thu, Nov 13, 2014 at 10:41 PM, Kapil Arya <[email protected]>
wrote:
> Hi Pushpendra,
>
> DMTCP ensures a consistent state of the distributed computations by the
> means of a centralized coordinator and network draining. You can find more
> information in this paper: http://dmtcp.sourceforge.net/papers/dmtcp.pdf.
>
> Please let me know if you need any more information.
>
> Kapil
>
>
> On Thu, Nov 13, 2014 at 12:09 AM, Pushpendra Singh <[email protected]>
> wrote:
>
>> Dear DMTCP team,
>>
>>
>>
>> Greetings! We are trying to use DMTCP for our work in coordinated
>> checkpointing. I have few queries in this regard with respect to
>> distributed applications (where multiple processes running on separate
>> nodes are communicating with each other and failure of any node means that
>> all nodes need to be restarted from the last checkpoint):
>>
>>
>>
>> 1. Does DMTCP checkpoints ensure consistent state, i.e. there are
>> no orphan messages?
>>
>> 2. If yes to the previous question then what is the underlying
>> algorithm for coordinated checkpointing in DMTCP? A link to corresponding
>> research paper will help.
>>
>> 3. If not then how can I integrate my “coordinated chekcpointing”
>> algorithm within DMTCP i.e. using DMTCP for mainly storing the checkpoint
>> while my algorithm ensures consistency.
>>
>>
>>
>>
>>
>> Thanks.
>>
>>
>>
>> Best,
>>
>> Pushpendra
>>
>> [image: logo]
>>
>> http://www.iiitd.edu.in/~pushpendra/
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Comprehensive Server Monitoring with Site24x7.
>> Monitor 10 servers for $9/Month.
>> Get alerted through email, SMS, voice calls or mobile push notifications.
>> Take corrective actions from your mobile device.
>>
>> http://pubads.g.doubleclick.net/gampad/clk?id=154624111&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Dmtcp-forum mailing list
>> [email protected]
>> https://lists.sourceforge.net/lists/listinfo/dmtcp-forum
>>
>>
>
------------------------------------------------------------------------------
Comprehensive Server Monitoring with Site24x7.
Monitor 10 servers for $9/Month.
Get alerted through email, SMS, voice calls or mobile push notifications.
Take corrective actions from your mobile device.
http://pubads.g.doubleclick.net/gampad/clk?id=154624111&iu=/4140/ostg.clktrk
_______________________________________________
Dmtcp-forum mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dmtcp-forum
