«

I have an application written in Perl/tk that will write a small file
(~128 bytes) to disk. At random intervals, the application is unable to
write ti the disk. No error message is given, the write file is
executed but the trap 'or die...' is not. The OS is RedHat 7.2.

I have checked the amount of disk space (~85% in use of a 40GB HDD) and
this tells me that it isn't a disk spave issue. If I shutdown and let
the server sit for about 20 minutes, then boot it back up, the system
will allow files to be written to disk once again. A couple of possible
explanations have come up:
1 - The OS is runnung out of file handles.
2 - There is a (possible) memory in perl 5.8.x

Has anyone seen this behaviour before?

TIA,

Joe

"" <> wrote:
> I have an application written in Perl/tk that will write a small file
> (~128 bytes) to disk. At random intervals, the application is unable to
> write ti the disk. No error message is given,

If no error messages are given, then how do you know that the application
is unable to write to the disk?

> the write file is
> executed but the trap 'or die...' is not. The OS is RedHat 7.2.

It is clear what you mean by this. Show us the code that this sentence
is supposed to describe.

Xho

--
-------------------- http://NewsReader.Com/ --------------------
Usenet Newsgroup Service $9.95/Month 30GB

"" <> wrote:
> I have an application written in Perl/tk that will write a small file
> (~128 bytes) to disk. At random intervals, the application is unable to
> write ti the disk. No error message is given,

If no error messages are given, then how do you know that the application
is unable to write to the disk?

> the write file is
> executed but the trap 'or die...' is not. The OS is RedHat 7.2.

It is not clear what you mean by this. Show us the code that this sentence
is supposed to describe.

Xho

--
-------------------- http://NewsReader.Com/ --------------------
Usenet Newsgroup Service $9.95/Month 30GB

To all who replied,

Thanks for your advice. It turns out their was a code condition that
was not being performed properly based on a decision tree I had
implemented. The crummy thing was it took 3 of us nearly a week to find
it.

The offending code has been removed and shelved for the time being,
pending a design rewwork.

Joe