More on SITEreader with CURRENT 20020311

Russ Allbery rra at stanford.edu
Sun Mar 17 07:52:16 UTC 2002



Jeffrey M Vinocur <jeff at litech.org> writes:
> On Thu, 14 Mar 2002, William F. Maton wrote:

>> Mar 14 20:10:05 mcast innd: SERVER internal SITEreader: litech:48:file
>> Mar 14 20:10:35 mcast last message repeated 866731 times

> Eugh.

> Russ...I forget why we wanted to know which channel it was that was
> causing the log messages.

Because it's selecting for readability, and whatever it is, it shouldn't
be doing that.  (SITEreader is the "null" reader function that INN uses.)

This looks like an output file for an outgoing feed.  This definitely
shouldn't be selecting for readability.  Furthermore, it shouldn't even be
in the set of fds that INN is *polling* for readability.  My guess is that
William's operating system decides that disk files will always select true
for readability; in fact, that's the documented behavior on Solaris.

The real question is how that channel ended up in the read mask.  Somehow,
RCHANadd is being called for it.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>


More information about the inn-bugs mailing list