My 9.5.1-P3 exit suddenly.

2009-08-11 Thread BBB Kee
Hi,

We have a intel solaris 9 and bind9.5.1-P3 inside it.  The named suddenly
stopped at this morning.  Here is it left:


..
11-Aug-2009 06:09:14.466 general: error: failed to start watching FD (512):
invalid file
11-Aug-2009 06:09:14.467 general: error: failed to start watching FD (512):
invalid file
11-Aug-2009 06:09:14.467 general: error: failed to start watching FD (512):
invalid file
11-Aug-2009 06:09:14.467 general: error: failed to start watching FD (512):
invalid file
11-Aug-2009 06:09:14.467 general: critical: socket.c:2413:
INSIST(!sock-pending_recv) failed
11-Aug-2009 06:09:14.468 general: critical: exiting (due to assertion
failure)


What is the problem?  Can I fix it?
___
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users

Re: My 9.5.1-P3 exit suddenly.

2009-08-11 Thread Dmitry Rybin

ulimit -a ? Looks like as max open file descriptor limit exceeded.

On FreeBSD/Linux boxes I use MONIT (http://mmonit.com/monit/) то check 
and restart bind.



BBB Kee wrote:

Hi,

We have a intel solaris 9 and bind9.5.1-P3 inside it.  The named suddenly
stopped at this morning.  Here is it left:


...
11-Aug-2009 06:09:14.466 general: error: failed to start watching FD 
(512): invalid file
11-Aug-2009 06:09:14.467 general: error: failed to start watching FD 
(512): invalid file
11-Aug-2009 06:09:14.467 general: error: failed to start watching FD 
(512): invalid file
11-Aug-2009 06:09:14.467 general: error: failed to start watching FD 
(512): invalid file
11-Aug-2009 06:09:14.467 general: critical: socket.c:2413: 
INSIST(!sock-pending_recv) failed
11-Aug-2009 06:09:14.468 general: critical: exiting (due to assertion 
failure)



What is the problem?  Can I fix it?




___
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users


___
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users

Re: My 9.5.1-P3 exit suddenly.

2009-08-11 Thread Gordon Ewasiuk
I was getting the same results on a couple of Solaris 9 on x86 servers.
During a code review, a coworker found some fairly major changes in how
BIND does things between 9.5.0-P2 and 9.5.1-P3.  

We had to backport just the security fix from 9.5.1-P3 to 9.5.0-P2 to
address the sock-pending_recv error.

 
On Tue, 2009-08-11 at 07:01 -0400, Dmitry Rybin wrote:
 ulimit -a ? Looks like as max open file descriptor limit exceeded.
 
 On FreeBSD/Linux boxes I use MONIT (http://mmonit.com/monit/) то check
 and restart bind.
 
 
 BBB Kee wrote:
  Hi,
 
  We have a intel solaris 9 and bind9.5.1-P3 inside it.  The named
 suddenly
  stopped at this morning.  Here is it left:
 
  
  ...
  11-Aug-2009 06:09:14.466 general: error: failed to start watching FD
  (512): invalid file
  11-Aug-2009 06:09:14.467 general: error: failed to start watching FD
  (512): invalid file
  11-Aug-2009 06:09:14.467 general: error: failed to start watching FD
  (512): invalid file
  11-Aug-2009 06:09:14.467 general: error: failed to start watching FD
  (512): invalid file
  11-Aug-2009 06:09:14.467 general: critical: socket.c:2413:
  INSIST(!sock-pending_recv) failed
  11-Aug-2009 06:09:14.468 general: critical: exiting (due to
 assertion
  failure)
 
 
  What is the problem?  Can I fix it?
 
 
 
 
 
  ___
  bind-users mailing list
  bind-users@lists.isc.org
  https://lists.isc.org/mailman/listinfo/bind-users
 
 ___
 bind-users mailing list
 bind-users@lists.isc.org
 https://lists.isc.org/mailman/listinfo/bind-users
 
 

___
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users

Re: My 9.5.1-P3 exit suddenly.

2009-08-11 Thread BBB Kee
Hi,

On Tue, Aug 11, 2009 at 7:01 PM, Dmitry Rybin kirg...@corbina.net wrote:

 ulimit -a ? Looks like as max open file descriptor limit exceeded.


time(seconds)unlimited
file(blocks) unlimited
data(kbytes) unlimited
stack(kbytes)8480
coredump(blocks) unlimited
nofiles(descriptors) 1024
vmemory(kbytes)  unlimited



 On FreeBSD/Linux boxes I use MONIT (http://mmonit.com/monit/) то check and
 restart bind.


We were running for years for old bind9 for a number of machines for years,
and haven't get this problem, and those are mostly for caching queries
only.  pfiles also only show 0-9,20-22,512,513,516,518,519,521 FD is using.
So I think it should
not hit FD limit, right?


Eric
___
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users

Re: My 9.5.1-P3 exit suddenly.

2009-08-11 Thread BBB Kee
On Tue, Aug 11, 2009 at 7:26 PM, Gordon Ewasiuk gewas...@above.net wrote:

 I was getting the same results on a couple of Solaris 9 on x86 servers.
 During a code review, a coworker found some fairly major changes in how
 BIND does things between 9.5.0-P2 and 9.5.1-P3.

 We had to backport just the security fix from 9.5.1-P3 to 9.5.0-P2 to
 address the sock-pending_recv error.



O...how?  I see lots of codes different between 9.5.0-P2 and 9.5.1-P3.  Is
it just copy the update.c from 9.5.1-P3 to 9.5.0-P2 and compile?  I just
make this changes and compile ok in 9.5.0-P2 and it seems it can prevent
from the current DOS attack also.

Eric
___
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users