Rootcint failure with batch submission

Hi rootcinters,

I am having trouble with rootcint when I submit jobs to a queue on a cray cluster. The jobs run fine when I run on the head node, or submit the job to an interactive node. However, when I submit the job to the batch queue I get the following error:

Error: /mnt/lustre_server/users/dcherdack/root/bin/rootcint: failed to open /tmp/rootcint_EicHPh.h, usage of external preprocessor by CINT is not optimal
Error: rootcint: failed to open /tmp/rootcint_lMl4YQ in main
Error in : Dictionary generation failed!
Info in : Invoking compiler to check macro’s validity
In file included from MCcheck_proxy_h_ACLiC.check.cxx:208:0:
/home/dcherdack/lustrefs/T2K//statcalc/ In member function ‘Double_t MCcheck_proxy::RecENue()’:
/home/dcherdack/lustrefs/T2K//statcalc/ error: ‘pow’ was not declared in this scope
/home/dcherdack/lustrefs/T2K//statcalc/ error: ‘sqrt’ was not declared in this scope

I have set the the TEMP env var to my own tmp dir, and can see files been written there as the job runs. I have also run the test script I found in another thread.

echo 'void script () { printf("we are fine\n"); }' > /tmp/script.C
root.exe -b -q -l /tmp/script.C
root.exe -b -q -l /tmp/script.C+

This ran fine interactively and when put into a bash script and run through the batch queue compute node, producing the output:

I saw in that same thread that this sort of error may be a red herring, but with no other information I am not sure what to try next.

I am running 5.28.00g

The cluster specifications are as follows:

[quote]System specifications
Cray model XT6m
Operational January 2011
Peak performance 12 teraflops
Dimensions: 7.5 ft. (h) x 2.0 ft. (w) x 4.5 ft. (d)
Computer partition
• 52 compute nodes
• 2 processors / node
• 104 AMD Magny Cours 64-bit 1.9 GHz total processors
• 12 cores / processor; 1,248 total cores
• 32 GB DDR3 ECC SDRAM / node; 1.664 TB total RAM
IO partition
• Lustre parallel file system
• 32 TB RAID array file system
• 600 MB / sec. IO bandwidth
• SeaStar2+ 2D torus topology
• 6.5 us latency
• 6.8 GB / sec. transfer speed
Cray Linux Environment (CLE) based on SUSE Linux
OpenMP and MPI2 message passing interface
• PGI, Cray, Pathscale
• C, C++, Fortran 77/90/95
CrayPAT performance analysis tools
Cray debugging utilities[/quote]

Please let me know if I can provide any other useful information.



Hi Dan,

Are the compiler tools installed on those batch nodes?