Re: sendmail 8.9.0

Bjarni R. Einarsson (bre@margmidlun.is)
Wed, 27 May 1998 13:26:24 +0000

> ftp://ftp.sendmail.org/pub/sendmail/KNOWNBUGS :
>
> * accept() problem on Linux.
> Apparently, the accept() in sendmail daemon loop can return ETIMEDOUT
> and cause sendmail to sleep for 5 seconds during which time no new
> connections will be accepted. An error is reported to syslog:
> Jun 9 17:14:12 hostname sendmail[207]: NOQUEUE: SYSERR(root):
> getrequests: accept: Connection timed out
> "Connection timed out" is not documented as a valid return from
> accept(2) and this is believed to be a bug in the Linux kernel.
>
This is a real problem for me - and I do believe it is a kernel bug.

Jaroslav Kysela posted a patch that *might* be addressing this same problem,
a few weeks ago. I will be testing the patch, w.r.t. this problem within
the next few days. The original message began like this:

> From: Jaroslav Kysela <perex@jcu.cz>
> Date: Tue, 21 Apr 1998 16:14:31 +0200 (MET DST)
> Subject: TCP accept bug in 2.0 and 2.0.34pre10 patch
>
> Hi!
>
> I found strange TCP accept bug in 2.0 kernels (include 2.0.34pre10).
> List of waiting TCP connections (socket -> receive_queue) can be at some
> times corrupted due to interrupts.
> This bug can produce some "zombie" sockets which aren't passed
> over accept() to user space, but leaves in kernel forever or until some
> timeout isn't expired.
[snip]

I'm not familiar with the kernel internals, but I imagine that the scenario
described here could cause some sort of resource exhaustion, which could then
cause the sendmail problem.. does that make sense?

-- 
Bjarni R. Einarsson
 bre@margmidlun.is               [ THIS SPACE INTENTIONALLY LEFT BLANK ]
 http://www.mmedia.is/~bre
 Juggler@IRCnet                  GOBBLE, GOBBLE, GOBBLE, GOBBLE, GOBBLE!
 


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu