Re: end-of-file on communication channel

From: BEERYR <Ron.Beery_at_DaytonOH.NCR.COM>
Date: 1996/12/06
Message-ID: <E2021p.C6F_at_ranger.daytonoh.ncr.com>#1/1


I had this mysterious problem a while back.

A long running batch process would fail every time about half way through. Turned on SQL trace and it ran fine, but very slow.

Turned out we would get a "Checkpoint not complete" message in the alert log and the db would freeze until the checkpoint completed, during which time the batch job failed with your error message.

Enlarging the redo logs and more of them along with making log_checkpoint_interval and log_checkpoint_timeout larger fixed the problem. (We already had checkpoint_process = true)

Of course there are many other causes to this error, this is one that had us and Oracle Support stumped for a week.

>=========="Steve Sherman", 12/5/96==========
>
>Has anyone encountered this error? We are having a hard time
 determining
>the sorts of things that would cause this. We suspect it might be a
>ramification of low system resources.
>
>If anyone can shed any light, I would realy appreciate it.
>
>Thanks,
> - Steve Sherman

Ron Beery
NCR
"Everything is computerized. What could possibly go wrong?" Received on Fri Dec 06 1996 - 00:00:00 CET

Original text of this message