[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: watchmalloc (Re: Corrupted messages...)

From: Dan Nelson
Subject: Re: watchmalloc (Re: Corrupted messages...)
Date: Thu, 12 Jun 2003 15:47:19 -0500
User-agent: Mutt/1.5.4i

In the last episode (Jun 12), Chuck Yerkes said:
> Well, murphy lives.
> Been running with watchmalloc for the last few days.
> Nothing amiss.  <sigh>.
> Load is up (from 0.11 to 4-5 on a 4-way machine).
> No cores yet.
> /me wonders if just the act of clearing memory it keeping
> the problem from happening.  Random thoughts move towards the
> problem being that some 8 bit data is being written to memory
> that's causing the some false End-Of-Something to be tripped.

Try this:  Log the first 20 bytes or so of the 'body' string just
before the call to smfi_replacebody.  If those come out garbled, the
problem is definitely in the milter and we can track it back to the
source.  If they're okay, then either smfi_replacebody is screwing up
or the body is getting corrupted inside sendmail after that.

You don't happen to have a Purify license? :)

        Dan Nelson

reply via email to

[Prev in Thread] Current Thread [Next in Thread]