Re: Broken Pipes in Server Log

Sonya Rikhtverchik (
Mon, 14 Jul 1997 10:49:59 -0400

Message-Id: <>
Subject: Re: Broken Pipes in Server Log
Date: Mon, 14 Jul 1997 10:49:59 -0400
From: Sonya Rikhtverchik <>

Date: Mon, 14 Jul 1997 10:50:04 -0400
From: Ben Engber <>
Message-Id: <>
Subject: Re: Broken Pipes in Server Log
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Md5: E1R05c5eBL789Kz/PsGdGQ==

We've been having this problem as well using the OpenMarket server 2.0 and 
fastCGI v 1.5.1 on Solaris 2.5.1 (programs are written in C++).  For now we're 
just restarting our servers every half hour, but it would be great to have an 
idea of what might be causing the problem.

Ben Engber

> 	I have seen a few e-mails in the archive from developers experiencing
> problems with broken pipe errors in the server logs. I am experiencing the
> same problem and I have been unable to find any e-mails that can expalain
> the problem or suggest a remedy. 
> 	The problem only occurs when I run the applications as a server 
> appclass (as opposed to an external app class). The likelihood of the
> problem occurring seems to be correllated to the load on the webserver or
> database server. Once the broken pipes begin, they remain broken and the
> webserver must be restarted to remedy the problem.
> 	As a temporary solution, I have made the fcgis an external app class. I
> would really like for the server to maintain the application in case the
> processees exit due to some fatal error. I am running the OpenMarket server
> 2.1 on Solaris 2.5.1 on an Ultra-2. If anyone has any information, I would
> really appreciate it. Thanks.
> Jon Perlow

- --
Ben Engber
Senior Programmer

------- End of Forwarded Message