As BrowserUk pointed out, if you let us know why the files failed to close, that would help. I'd like to add that there is no "best way to manage" this problem. It depends upon what you are trying to do, why you are trying to do it, what happens if you don't do it, etc.

Aside answering the questions above, here are some others:

• If you use source control, check what changes have occurred to the offending programs when the failures started to occur. Rarely do problems just "start happening".
• Is this really a cgi environment? Any chance that it's mod_perl, ISAPI, or something else that might have a bearing on the problem?
• Can you show us any of the offending code?
• You mentioned that this is a recent problem. Has anything on your box changed? Is there any scheduled maintenance you skipped? I know the latter question seems weird, but if you don't know what is causing a problem, then everything is suspect.

Cheers,
Ovid

Join the Perlmonks Setiathome Group or just click on the the link and check out our stats.

My favorite guess is that the handle is already closed. A close call can also expose errors from buffered prints.

Check your error logs.

After Compline,
Zaxo

A close call can also expose errors from buffered prints.

This can happen when disk usage limits are strictly enforced. (I.e., "You can write only this much; you can't write no more.") And, since people almost never check errors on writes, errors go undetected until close.

It would be easier to make suggestions if you told us what reason(s) Perl identifies (in $!) for the failures to close?

Yah, I shoulda printed/saved that... Now I'm waiting for the .1% of the time when it screws up.

The error log (just one file - .6GB - for the entire server) doesn't provide any insight.

--lnl