Thanks to:
John P Speno [speno_at_isc.upenn.edu]
David J. DeWolfe [sxdjd_at_java.sois.alaska.edu]
Attached below is my original question. John hit the nail on the head when
he suggested that the ulimit datasize was too small. We were set at
131073KB, I raised it to 500000KB and the compile worked. (High energy
physics research requires huge programs to analyze vast amounts of data.)
I sent out the UPDATE to the problem because being a "ksh" kind of guy I
tried to increase his datasize with the "ulimit -d 500000" command. This
user runs a "tcsh", a derivative of the "csh". To make a long story short
"ulimit" doesn't work in "csh". In "csh" the command is "limit datasize
500000". I found this in "The UNIX C Shell Field Guide" - Prentice/Hall.
Putting the proper command in the users ".cshrc" file worked like a charm.
Although the compile problem is solved, I now have another problem. I have
other users, many in fact, that will also be compiling large binaries which
will exceed the default 131072MB datasize. How can I raise that default
datasize globally? Where is it located? What command do I have to use? Is it
like AIX where I can just edit a file? Maybe I should submit this question
separately?
Anyway this is a great list, once again the list comes through.
Jim Fitzmaurice
jpfitz_at_fnal.gov
UNIX is very user friendly, It's just very particular about who it makes
friends with.
---[Original
Question]----------------------------------------------------------
Managers,
Below is an excerpt from a compile. The engineer compiling the code thinks
a kernel parameter need to be adjusted to compile the code, since he is able
to compile it in a SGI machine with a similar compiling package.
I'm at a loss to determine which kernal parameter may alleviate the error.
Any help would be appreciated. I'm running 4.0d PK#3 on a 4100.
#
#---------------------------------------------------------------------------
--
#
--> (GNUmakefile.main) _bin_ Fri Jan 7 23:05:52 CST 2000 jetanalyze
(jetanalyze)
--> (ctest_sequencer.mk) _bin_ Fri Jan 7 23:05:53 CST 2000 jetanalyze
(jetanalyze/src)
--> (ctest_sequencer.mk) _bin_ Fri Jan 7 23:05:58 CST 2000 jetanalyze
(jetanalyze/bin)
--> (ctest_rules.mk; __jetanalyze__) Compiling binary JetAnalyze_x.
ld_mmap: Could not malloc 8609792. bytes, for
/usr/products/root/v2_23_09/OSF1/lib/libCore.so, didnt attempt to mmap
ld_mmap: Could not malloc 8609792. bytes, for
/usr/products/root/v2_23_09/OSF1/lib/libCore.so, didnt attempt to mmap
file file_offset / flen -- pg bndry file
offst --> last file offst mmapd
mmap usage info not available; max of 160. slots
Fatal error in: /usr/lib/cmplrs/cc/ld IOT/Abort trap - core dumped
gmake[2]: ***
[/d0dist/dist/releases/t00.72.00/bin/OSF1V4-KCC_3_3/JetAnalyze_x] Error 127
gmake[1]: *** [bin.bin] Error 2
gmake: *** [jetanalyze.bin] Error 2
#
Jim Fitzmaurice
jpfitz_at_fnal.gov
UNIX is very user friendly, It's just very particular about who it makes
friends with.
Received on Fri Jan 21 2000 - 19:26:57 NZDT