BIND 8.2.1 Hangs - NDC fails

Mark_Andrews at isc.org Mark_Andrews at isc.org
Mon Sep 13 21:54:36 UTC 1999


	Apply the following patch.

	Mark

Index: src/bin/named/ns_main.c
===================================================================
RCS file: /proj/cvs/isc/bind/src/bin/named/ns_main.c,v
retrieving revision 8.110
retrieving revision 8.111
diff -c -r8.110 -r8.111
*** ns_main.c	1999/09/04 22:06:13	8.110
--- ns_main.c	1999/09/13 06:33:46	8.111
***************
*** 1759,1765 ****
   */
  int
  sq_openw(struct qstream *qs, int buflen) {
! #ifdef SO_LINGER	/* XXX */
  	static const struct linger ll = { 1, 120 };
  #endif
  
--- 1759,1765 ----
   */
  int
  sq_openw(struct qstream *qs, int buflen) {
! #ifdef DO_SO_LINGER	/* XXX */
  	static const struct linger ll = { 1, 120 };
  #endif
  
***************
*** 1770,1776 ****
  	qs->s_wbuf_send = qs->s_wbuf;
  	qs->s_wbuf_free = qs->s_wbuf;
  	qs->s_wbuf_end = qs->s_wbuf + buflen;
! #ifdef SO_LINGER	/* XXX */
  	/* kernels that map pages for IO end up failing if the pipe is full
  	 * at exit and we take away the final buffer.  this is really a kernel
  	 * bug but it's harmless on systems that are not broken, so...
--- 1770,1776 ----
  	qs->s_wbuf_send = qs->s_wbuf;
  	qs->s_wbuf_free = qs->s_wbuf;
  	qs->s_wbuf_end = qs->s_wbuf + buflen;
! #ifdef DO_SO_LINGER	/* XXX */
  	/* kernels that map pages for IO end up failing if the pipe is full
  	 * at exit and we take away the final buffer.  this is really a kernel
  	 * bug but it's harmless on systems that are not broken, so...

> Barrett Richardson wrote:
> 
> > On 7 Sep 1999, Barry Margolin wrote:
> >
> > > In article <Pine.LNX.4.10.9909080037260.15487-100000 at brodie.two-four.com>
> ,
> > > Adam D. Burstein <adam at register.com> wrote:
> > > >We are also getting a lot of unauthorized updates (which I would like to
> > >
> > > Others have reported named hanging in conjunction with lots of these.  It
> > > sounds like there may be a bug, but it hasn't been tracked down.
> > >
> >
> > I've been having the same problem for about a week now. I've got
> > a script running that does a 'ndc reload' if nslookup doesn't get
> > a response for a few seconds. I've noticed some patterns around
> > my reloads in the logfiles.
> >
> >   1. The logfile is quiet for serveral minutes. Normally
> >      I get around 75 messages per minute if logging noise
> >      like Lameserver and cname.
> >
> >    2. Then I get several messages in the same second like
> >       ns_req: sendto(): Connection refused
> >       and in the mix, still the same second, I get occasional
> >       stream_getlen(): Broken pipe
> >       messages intermingled with the ns_req messages.
> >
> >     3. Then usually a second later I see the reload message
> >        in the logfile.
> >
> > -
> >
> > Barrett
> 
> I have the same problem here at work. Any chance you could post that script t
> o
> the newsgroupor make it available?
> 
> - Eric Molitor
> 
> 
> 
--
Mark Andrews, Internet Software Consortium
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: marka at isc.org


More information about the bind-users mailing list