Re: a question about BLCR

From: Sergio Díaz (sdiaz_at_cesga.es)
Date: Mon Jun 01 2009 - 03:57:03 PDT

  • Next message: Peter Elias: "Re: math.ct failure"
    Hi,
    
    Sorry, but I'm working with BLCR and I don't know if there is other 
    application to support network comunications... You can take a look to 
    this tool but I don't know if it works and how it works because I 
    haven't tested it.
    
    http://hpc.pnl.gov/sft/tick.html
    
    Regards,
    Sergio
    
    
    shuxi liu escribió:
    > hello Paul and Sergio,
    >    thanks for all your help.In my opinino,maybe I can't use BLCR to do 
    > checkpoint t abuot Apache.
    >   But nowadys I do some research in Software fault-tolerance.I do need 
    > to do checkpoint t abuot Apache.
    >  Can you  recommend some software to make it?
    >  
    > thank you for your help.
    >  
    >  
    >                                                        Jack
    >
    > 2009/5/27 Paul H. Hargrove <PHHargrove_at_lbl_dot_gov 
    > <mailto:PHHargrove_at_lbl_dot_gov>>
    >
    >     Jack,
    >
    >
    >     There are at least 2 things that come to mind for the error you report
    >     1) As Sergio has suggested, you may have an LD_LIBRARY_PATH that
    >     is pointing only to 32- or 64-bit libraries when apache needs the
    >     one not in LS_LIBARARY_PATH.
    >
    >     2) If the apache binary is setuid or setgid, then I beleive the
    >     system will refuse/ignore any LD_PRELOAD settings as a security
    >     measure.  If that is the case, then there is probably no way to
    >     get cr_run and setuid/setgid to work together.
    >
    >     -Paul
    >
    >
    >     Sergio Díaz wrote:
    >
    >         Hello,
    >
    >         This error seems that you are trying to run an application
    >         compiled in other architecture (64 or 32 bits).
    >         Maybe Paul can give you more clues.
    >
    >         Regards,
    >         Sergio
    >
    >
    >         shuxi liu escribió:
    >
    >             hello,
    >               thanks for your reply.As you know,I have done the
    >             cr_checkpoint and cr_restart in example of counting.
    >              but when i use cr_cun,also some errors.
    >             [root@fedora10 builddir]# cr_run
    >             /root/blcr-0.8.1/builddir/examples/counting/counting
    >             ERROR: ld.so: object 'libcr_run.so.0' from LD_PRELOAD
    >             cannot be preloaded: ignored.
    >             can you tell me how to make it,thank you very much
    >                           jack
    >              
    >
    >              2009/5/25 Sergio Díaz <sdiaz@cesga.es
    >             <mailto:sdiaz@cesga.es> <mailto:sdiaz@cesga.es
    >             <mailto:sdiaz@cesga.es>>>
    >
    >                Hi,
    >
    >                I don't know why are you trying to do a checkpoint to
    >             apache....
    >                and if it is possible to do a checkpoint to a daemon,
    >             but to get a
    >                successful checkpoint, you have to run the application
    >             with the
    >                command cr_run. For example: "cr_run ./sleep.sh" then
    >             you can do
    >                 cr_checkpoint.
    >                Also, you have to load the checkpoint modules (blcr.ko and
    >                blcr_import.ko). If not, the error is the same.
    >
    >
    >                Regards,
    >                Sergio
    >
    >
    >                shuxi liu escribió:
    >
    >                    hello sir or madam,
    >                        I have used blcr-0.8.1 in
    >             2.6.27.5-117.fc10.i686,and i
    >                    have done the example of counting.
    >                        I want to make checkpoint to apache, but i have
    >             some problem.
    >                        httpd pid is 3170
    >                      [root@fedora10 blcr-0.8.1]# cr_checkpoint 3170
    >                      Checkpoint failed: support missing from kernel
    >                      i don't know how to make checkpoint to apache.i
    >             need your help.
    >                      i look forward to your reply.thank you very much.
    >                                                     turely yours      
    >                                                                      
    >              Jack
    >                      
    >
    >                --     Sergio Díaz Montes
    >                Centro de Supercomputacion de Galicia
    >                Avda. de Vigo. s/n (Campus Sur) 15706 Santiago de
    >             Compostela (Spain)
    >                Tel: +34 981 56 98 10 ; Fax: +34 981 59 46 16
    >                email: sdiaz@cesga.es <mailto:sdiaz@cesga.es>
    >             <mailto:sdiaz@cesga.es <mailto:sdiaz@cesga.es>> ;
    >             http://www.cesga.es/
    >                ------------------------------------------------
    >
    >
    >
    >
    >
    >
    >     -- 
    >     Paul H. Hargrove                          PHHargrove_at_lbl_dot_gov
    >     <mailto:PHHargrove_at_lbl_dot_gov>
    >     Future Technologies Group
    >     HPC Research Department                   Tel: +1-510-495-2352
    >     Lawrence Berkeley National Laboratory     Fax: +1-510-486-6900
    >
    >
    
    
    -- 
    Sergio Díaz Montes
    Centro de Supercomputacion de Galicia
    Avda. de Vigo. s/n (Campus Sur) 15706 Santiago de Compostela (Spain)
    Tel: +34 981 56 98 10 ; Fax: +34 981 59 46 16
    email: sdiaz@cesga.es ; http://www.cesga.es/
    ------------------------------------------------ 
    

  • Next message: Peter Elias: "Re: math.ct failure"