“Memory has been corrupted” error on linux

Guest wrote on Jun 12, 2006

Hi Iraf folks,

Any suggestions on how to beat this problem? When I run this command:

crosscor input1=j8rq22maq_single_sci.fits input2=j8rq22maq_single_sci.fits
output=xcor_j8rq22maq_j8rq22maq.fits coord_shift=no center=yes pad=no

I get this output:
input1 r: j8rq22maq_single_sci.fits
input1 i: [none]
input2 r: j8rq22maq_single_sci.fits
input2 i: [none]
output r: xcor_j8rq22maq_j8rq22maq.fits
output i: [none]
PANIC in `/iraf/stsdas/bin.redhat/x_fourier.e': Memory has been corrupted

I'm running iraf 2.12.2a-EXPORT on fedora core 3, kernel
2.6.11-13_FC3aslsmp, gcc 3.4.2. I read the exec-shield discussion on the
release notes, but the file /proc/sys/kernel/exec-shield doesn't exist on
my system.

This exact command has worked on this machine in the past.

Thanks!
Jane R.

_____________________________________________________________________

Dr. Jane Rigby jrigby@as.arizona.edu
Postdoctoral Fellow MIPS Group
Steward Observatory The University of Arizona
520-621-2589 (work) 520-668-4290 (cell)

"I may be vegetarian, but I still keep my eyes in the front of my head."
_____________________________________________________________________

Post generated using Mail2Forum (http://www.mail2forum.com)

---

Mike Fitzpatrick wrote on Jun 12, 2006

Memory corrupted errors are usually coding bugs, sometimes data-specific. Since you say the task worked before we can probably rule out problems with the STSDAS binaries but if you're using an older version of STSDAS it might be worth updating (or checking) to see if this is a fixed known bug. Otherwise, a message to help@stsci.edu may be your best bet. They may need the data in question to reproduce the problem.

Note that simply doing an "unlearn" on the task never hursts either, there are sometimes conflicting parameter settings from earlier runs that could trigger the error as well. A quick forum search shows that hasn't been reported earlier.

Cheers,
-Mike

---

Last post on Jun 12, 2006