Re: [Condor-users] Does the standard universe really work on x86_64?

[Mark Calleja <M.Calleja@xxxxxxxxxxxxxxx>]

Seriously, can no one help out with our Standard universe + Condor 6.8.2 
+ x86_64 Linux problem? Go on, send some early Christmas cheer this way.

Mark

Mark Calleja wrote:
> We're having very little joy getting the standard universe to work with 
> the x86_64 build for Condor 6.8.2, and we'd be ecstatically pleased if 
> *anyone* could shed some light on the matter. First, here are the 
> details of the Linux platform we're testing on:
>
> Machine OS: Suse Enterprise Server 10
> uname -a: Linux <hostname> 2.6.16.21-0.25-default #1 Tue Sep 19 07:26:15 
> UTC 2006 x86_64 x86_64 x86_64 GNU/Linux
>  From /proc/cpuinfo: model name : AMD Athlon(tm) 64 Processor 3200+
>
> As suggested on the Condor Download page, we've installed the 
> dynamically-linked version of the RHE3 Opteron 6.8.2 build. The software 
> we've tested is a C++ application that works fine both with the native 
> g++ v4.1.0 on the machine as well as g++ v3.4.6 that we also installed. 
> Now for condor_compiling: condor_compile builds and links "fine" with 
> g++ v3.4.6 but fails with v 4.1.0, though the latter can be made to link 
> if  -fno-threadsafe-statics is added to the compilation flags. If we now 
> try to run either executable at the command line (never mind submitting 
> as a Condor job) then both fail with a segment violation. Here's what 
> gdb has to say about it:
>
> ===
> Starting program: 
> /home/Condor/mark/test/diffmc/diffmc-standard-x86_64-v4.1.0 NVT_PARAM 
> CONFIG
> Condor: Notice: Will checkpoint to 
> /home/Condor/mark/test/diffmc/diffmc-standard-x86_64-v4.1.0.ckpt
> Condor: Notice: Remote system calls disabled.
>
> Program received signal SIGSEGV, Segmentation fault.
> 0x00000000004adeeb in std::locale::operator= ()
> ===
>
> I'm guessing that this is a dynamic library incompatibility problem, but 
> how do we get out of this pickle? What x86_64 linux distros are people 
> successfully using with the standard universe? For what it's worth, we 
> can get the standard universe to work on i386 platforms OK.
>
> Thanks,
> Mark
>
>