Dear all,
many thanks for your comments.
My OS is Windows 7.
I will try to describe the program more in detail:
- The overall idea is to simulate a production planing done in rolling schedule
- The main routine is in Java. Here I generate test data that I export into a couple of.csv.-files. There is only a (public static void) main routine in Java, no other thread. For each csv.-file I do a .close Statement after finishing the writing.
- After generating the .csv-files the routine calls a GLPK model (see the source code of my first post)
- The GLPK model reads in the .csv-files and exports the solution into other .csv. files
- Then the Java program reads in these csv. files, does some operations on the data and generates new output .csv files that are read in by GLPK and so on.
After some iterations I get the mentioned error. Not always for the same .csv-file, sometimes also for a different one, and in a different iteration. This would indicate that Nigel is right. However, I could not detect other processes than my Java main routine (and other open programs like Eclispe) using the tool suggsted by Nigel.
Do you have some further ideas on that?
Martin
Let us assume the OS is Windows in which case 113 means the system has
run out of file handles. Normally you can open a file to read even if
something else has it open for writing. I would suggest installing
http://technet.microsoft.com/en-gb/sysinternals/bb896653 and observe
which processes are opening so many files then fixing the code.
--
Nigel Galloway
address@hidden
On Tue, Jan 6, 2015, at 12:21 PM, Michael Hennebry wrote:
On Tue, 6 Jan 2015, Martin Albrecht wrote:
I am running a GLPK model that is sequentially started from a frame program
in Java.
In this contex, what is a frame program?
What OS are you running?
Unfortunately, I get sometime error when opening the .csv files.
Here is an example:
/
csv_driver: unable to open BOM.csv - Permission denied
allocation.mod:113: error on opening table tab_BOM
glp_mpl_build_prob: invalid call sequence
Error detected in file ..\src\glpapi14.c at line 93/
I do not get this error at each call of the Java model (one error after
approx. 50 calls). Moreover, these errors do not occur in a deterministic
manner, they cannot be reproduced. I each run of the Java program they occur
at another place.
Looks like a thread issue to me.
Do the open and close occur in different threads?
Perhaps the open is attempted before the close has gone through.
--
Michael address@hidden
"SCSI is NOT magic. There are *fundamental technical
reasons* why it is necessary to sacrifice a young
goat to your SCSI chain now and then." -- John Woods
_______________________________________________
Help-glpk mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/help-glpk