Re: [Dmtcp-forum] Problem with forked checkpointing

[Kapil A. <ka...@cc...>]

Hello Jake,

Thanks for the bug report. I will look into it and write you back.

Thanks,
-Kapil

On Fri, Oct 21, 2011 at 5:51 PM, Gamage, Jacob J
<jac...@in...>wrote:

> Hi there,
>
> First off - thanks very much for putting together such an interesting and
> useful project!
>
> I'm having some issues with the "forked checkpointing" feature and multiple
> checkpoints.  I'm using DMTCP 1.2.3 on a SLES10 system with kernel version
> 2.1.16-60.
>
> DMTCP is configured like this:
> ./configure --enable-unique-checkpoint-filenames
> --enable-forked-checkpointing CPPFLAGS="-DDEBUG" CXXFLAGS="-O0 -g3"
> --enable-debug=full
>
> I made a simple test program:
>
> #include "dmtcpaware.h"
> #include <unistd.h>
> #include <iostream>
>
> using namespace std;
>
> int main() {
>        int count = 0;
>        cout<<"Main function"<<endl;
>        cout<<"DMTCP available:"<<dmtcpIsEnabled()<<endl;
>        while(count < 60) {
>                cout<<"count is "<<count<<endl;
>        if(dmtcpIsEnabled() && ((count == 2) || (count == 22))) {
>            cout<<"Checkpointing."<<endl;
>            dmtcpCheckpoint();
>        }
>                count++;
>        sleep(1);
>        }
>
>        cout<<"End."<<endl;
>        return 0;
>
> }
>
>
> When I run this, the second checkpoint fails with this error:
>
> [8958] ERROR at connectionmanager.cpp:122 in retrieve; REASON='JASSERT(i !=
> _table.end()) failed'
>     fd = 4
>     device = pipe:[176227488]
>     _table.size() = 2
>
> Here is the "Listings FDs" printed out:
> [8958] TRACE at connectionmanager.cpp:444 in dbgSpamFds; REASON='Listings
> FDs...'
>     out.str() =
>        fd -> device             -> inTable
>        0 -> pts:/dev/pts/20 inTable=1
>        1 -> pts:/dev/pts/20 inTable=1
>        2 -> pts:/dev/pts/20 inTable=1
>        3 -> file[3]:/tmp/dmtcp.1QmyuR (deleted) inTable=0
>        4 -> pipe:[176227488] inTable=0
>        5 -> pipe:[176227488] inTable=0
>
> I'm guessing that it has a problem with the pipe that the first checkpoint
> created to gzip the output file.  The test works if I disable forked
> checkpointing.
>
> Can you reproduce this bug?
>
> Thanks very much,
> Jake Gamage
>
>
>
>
> ------------------------------------------------------------------------------
> The demand for IT networking professionals continues to grow, and the
> demand for specialized networking skills is growing even more rapidly.
> Take a complimentary Learning@Cisco Self-Assessment and learn
> about Cisco certifications, training, and career opportunities.
> http://p.sf.net/sfu/cisco-dev2dev
> _______________________________________________
> Dmtcp-forum mailing list
> Dmt...@li...
> https://lists.sourceforge.net/lists/listinfo/dmtcp-forum
>