inn 2.4.3 segfaults in certain incoming.conf situations

Marc Haber mh+inn-workers at zugschlus.de
Sat Jun 23 17:15:25 UTC 2007


On Sat, Jun 23, 2007 at 09:55:17AM -0400, Bill Davidsen wrote:
> Whenever I see "sometimes" regarding INN, I think hardware, because INN 
> is pretty deterministic, and blowing up every now and then on the same 
> input data is unlikely. Therefore I would look at hardware to see if 
> something is on the edge of reliability and failing under the load of 
> startup activity.

Nack, the bug is reproducible on a reference box. I think that the
same failure on two boxes nearly rules out hardware since both boxes
need to be defective identically. Unlikely, IMO.

> The other possibility is that there is an address (or several) which 
> produce DNS misbehavior. INN having some lightly traveled code paths in 
> the "DNS did something unexpected" area seems at least possible if not 
> likely. If you can monitor the network traffic between your DNS server 
> and INN at startup you might gain some useful insight.

I can do that, yes. Can you explain how a DNS issue can happen when I
exchange two peers in the incoming.conf?

> The code doesn't change, the incoming.conf doesn't change, therefore 
> something else does. It could be hardware related, it could be timing 
> related (unlikely), or it could be DNS. DNS is probably the easiest 
> thing to check.

Will do.

$ sudo tcpdump -i eth0 -s 2000 -w tcpdump.dns tcp port 53 or udp port 53
tcpdump: listening on eth0, link-type EN10MB (Ethernet), capture size 2000 bytes

Greetings
Marc

-- 
-----------------------------------------------------------------------------
Marc Haber         | "I don't trust Computers. They | Mailadresse im Header
Mannheim, Germany  |  lose things."    Winona Ryder | Fon: *49 621 72739834
Nordisch by Nature |  How to make an American Quilt | Fax: *49 3221 2323190


More information about the inn-workers mailing list