> Hello!
>
> > So it looks like sk_zapped is set to 0 and the timer has stopped.
>
> Damn.
>
> OK, Chingiskhan used to say: "The more busy grass is the easier to scyze" 8)
Well, I decided to try out 2.2.4 on that box and I -think- it got much
worse that with Andi's patch. With 24 hours of uptime I had a 112 stuck
sockets with wu's limit set to 180 this effectively crippled the server :<
Andi, is it possible that you were on the right track but there is anothe
condition? Can I get you guys any more information?
If I don't hear anything I'll just patch wu to have it's own connect
timeout, but that really doesn't help fix the kernel problem :<
Also, with plain 2.2.4 I am getting -lots- of these messages:
*** tcp.c:tcp_data bug acked < copied
*** tcp.c:tcp_data bug acked < copied
It might be that Andi's patch to connect() had some effect on this message
or that network conditions changed again.. After Andi sent me the first
patch for 2.2.3 these messages disappeared for a week.
Still getting lots of the
sending pkt_too_big to self
Which I understand to be a harmless message..
Thanks,
Jason
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/