Re: brat not compiling with RH7.1

From: Christian Holm Christensen (cholm@hehi03.nbi.dk)
Date: Fri Jan 04 2002 - 13:24:38 EST

  • Next message: Stephen J. Sanders: "need help with new file set stuff"

    Hi Hiro, 
    
    On Fri, 4 Jan 2002 07:43:18 -0600 (CST)
    Hironori Ito <hito@students.phsx.ukans.edu> wrote
    concerning "Re: brat not compiling with RH7.1":
    > 	Ok.  BRAT does work with RH7.1 in my pc w/ my own root w/o
    > cross compiler (I do not have egcs and egcs++ option).  As people
    > suggested, there must be a confict in root or linker or something else in
    > pii machine.  Or well, I guess I do not care as long as there is one way
    > to compile brat.  But, we should keep only one version of RH since it is
    > very annoying.
    
    Unfortunately, that is not an option.  We are more or less bound by
    RCF to at least have a few machines that are similar to the rcas and
    rcrs nodes.  At the same time, we sometimes need various features from
    never releases of Red Hat than those in the rcas configuration. 
    
    Now, that's just the situation at BNL.  There are some 12 institutes
    in the collaboration, and you can bet your arse that each of them has
    at least one configuration that is different from the rcas one.  
    
    No, the proper way to get rid of the problem you had, is to make the
    Red Hat 7.x machines identify themselves to AFS differently than the
    Red Hat 6.x machines.   The identification is reported by doing 
    
      fs sysname 
    
    Here's a list of some machines, their characteristics, and output of
    the above command: 
    
      host           | RH  | Kernel   | GCC     | Sysname 
      ---------------+-----+----------+---------+--------------
      rcas0014       | 6.1 | 2.2.18   | 2.91.66 | i386_redhat61
      rplay2         | 6.1 | 2.2.18   | 2.91.66 | i386_redhat61 
      pii4           | 6.2 | 2.2.17   | 2.91.66 | i386_redhat61
      pii10          | 7.1 | 2.4.2    | 2.96    | i386_redhat61
      hilux12.nbi.dk | 6.2 | 2.2.16   | 2.91.66 | i386_linux22
      ekman.nbi.dk   | 7.1 | 2.4.3    | 2.96    | i386_linux24
      hehi01.cern.ch | 6.1 | 2.2.12   | 2.91.66 | i386_linux22
      
    As you can see, very different systems, espcially rcas, pii4, and
    pii10, but they all identify themselves as i386_redhat61, which means
    that 
    
      /afs/rhic/opt/@sys/brahms 
    
    on all of rcas, pii4, and pii10, evaluate to 
    
      /afs/rhic/opt/i386_redhat61/brahms 
    
    and hence 
    
      /afs/rhic/opt/brahms 
    
    all point to the same directory, namely the one above.  This is
    obviously wrong - pii10 is not at all like the other machines, and
    pii4 is not Red Hat 6.1.  If someone could fix it so that the Red Hat
    7.1 machines identifies themselves as something like i386_redhat71,
    then you would have these problems, 'cause then on these machines, 
    
      /afs/rhic/opt/brahms 
    
    would resolve to 
    
      /afs/rhic/opt/i386_redhat71/brahms 
    
    and software installed on that filesystem would be compatible with Red
    Hat 7.1.  This is the force of AFS, and not utilising it properly is
    downright stupid. 
    
    Yours, 
    
    Christian Holm Christensen -------------------------------------------
    Address: Sankt Hansgade 23, 1. th.           Phone:  (+45) 35 35 96 91 
             DK-2200 Copenhagen N                Cell:   (+45) 28 82 16 23
             Denmark                             Office: (+45) 353  25 305 
    Email:   cholm@nbi.dk                        Web:    www.nbi.dk/~cholm
    



    This archive was generated by hypermail 2b30 : Fri Jan 04 2002 - 13:25:37 EST