ERROR: Seek error on file
corrado wrote on Apr 05, 2015
Hello all,
the lastest version of IRAF I installed (2.16) seems not to take lists of spectra, the problem is not bound to one particular task. For instance, by using scopy on a list of spectra @list.fits I get the error
ERROR: Seek error on file (giskard!/home/corrado/spectra/list.fits)
"verbose=verbose)"
line 29: onedspec$scopy.cl
called as: `scopy (mode=h)'
I get a similar error if I call wspectext, dispcor or another task. What is wrong?
The same list of spectra work without errors on another machine, so I suppose there is some wrong on the settings of IRAF on my machine.
Thank you for any suggestion/help you can give me.
Cheers,
Corrado
Mike Fitzpatrick wrote on Apr 06, 2015
I can think of three possible causes:
1) If the file list is especially long (e.g. hundreds of file paths) then you may be seeing a bug in the new image template code, however this normally shows up as a segfault in the task. The workaround is to do
cl> reset use_new_imt = no
before running the task. You can make this change permanent by editing the hlib$zzsetenv.def file or putting this statement in your login.cl before the final 'keep'
2) The FITS file itself is actually corrupt. Can you run a task on just the failing file? Is it always this file name, or this position in the list?
3) The file is > 2GB in size or on a 64-bit filesystem and you are using a 32-bit IRAF. If true, there is no workaround other than to use a 64-bit version of IRAF.
corrado wrote on Apr 09, 2015
Thank you for your answer!
your points 1) and 2) does not apply because the list is not too long and the fits files are not corrupted. In fact I could process them with IRAF on another machine. The machine that gives the error can process the fits files one by one correctly, but when I give the list of the same spectra, it gives the error.
About point 3) the files are not so big and I'm pretty sure I installed the 64-bit IRAF, since the machine is a 64-bit one.
still looking for a solution...
corrado wrote on Apr 09, 2015
I correct myself!
I tried to set the "cl> reset use_new_imt = no" that you suggested, and it works!
Thank you very much!
Cheers,
Corrado
Last post on Apr 09, 2015