httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Justin Erenkrantz <jerenkra...@apache.org>
Subject Re: [PATCH] Alerting when fnctl is going bad
Date Thu, 26 Sep 2002 01:25:05 GMT
On Thu, Sep 26, 2002 at 03:00:56AM +0200, Dirk-Willem van Gulik wrote:
> Well the numbers in that patch seem to work on sol/bsd/linux since 1.3.9
> under just about any set of circumstances.
> 
> So let's think out of the box then - how can we prevent apache from
> spinning *forever* in that while loop when something has gone astray ?

If I'm understanding you correctly, you want to detect when EINTR
really means "The server is hosed."  What is the correlation?  I
think we should only interpret EINTR as 'try again.'  

If there's an error (i.e. not EINTR) returned by fcntl(), then we'll
handle it accordingly and exit.  But, I just don't see how EINTR (no
matter how many) is indicative of a fatal error.  -- justin

Mime
View raw message