DNS "stalls" on Solaris2.6/BIND8

Mark_Andrews at isc.org Mark_Andrews at isc.org
Mon Oct 18 05:11:53 UTC 1999


	Apply the following.

*** src/bin/named/ns_main.c.orig	Wed Apr 29 05:17:46 1998
--- src/bin/named/ns_main.c	Tue Oct  5 12:12:04 1999
***************
*** 1655,1661 ****
   */
  int
  sq_openw(struct qstream *qs, int buflen) {
! #ifdef SO_LINGER	/* XXX */
  	static const struct linger ll = { 1, 120 };
  #endif
  
--- 1655,1661 ----
   */
  int
  sq_openw(struct qstream *qs, int buflen) {
! #ifdef DO_SO_LINGER	/* XXX */
  	static const struct linger ll = { 1, 120 };
  #endif
  
***************
*** 1666,1672 ****
  	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...
--- 1666,1672 ----
  	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...
> Hi All
> 
> I'm experiencing an intermittent fault on a new name server I'm setting
> up under Solaris.
> 
> When resolving names (locally set up or remote, authoritative or not, in
> cache or not) it sometimes takes several minutes to respond. Then the
> next second the same lookup works instantly. Then at some random
> interval a particular address mike not resolve again, then it works fine
> again etc.etc. This can be from a telnet session on the box itself
> (e.g.. ping host.domain.com or a remote PC using it as a NS)
> 
> The problem can be seen on an Ultra Sparc running Solaris 2.6 (all the
> latest Recommended patches applied). Other 4.9.7, 8.1.2 or 8.2.1
> installations on similar boxes with similar configs etc. work fine. The
> box has heaps of RAM, disk and CPU to spare though it is doing plenty of
> other things (very fast!)
> 
> I suspect a problem with either the resolver libraries or memory but I'm
> really guessing. The only clues I have is that nslookup works fine all
> the time and named is not growing large in RAM like it used to under
> BIND 4.9.7 as the cache built up.
> 
> Anyone who has some experience, clues, hunches, good tools, similar
> horror stories or the meaning of life to share, your input would be
> appreciated.
> 
> Cheers
> 
> 
> =====================================================================
> Andrew Sinclair                 Ph  (07) 4033 2311
> Technical Director              Fax (07) 4033 6022
> Internet North                  email: andrew at internetnorth.com.au
> A division of EISA Limited      www.internetnorth.com.au
> ISP for Cairns and North Qld    PO Box 12272D Cairns Qld 4870
> =====================================================================
> 
--
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