On RHEL5 with kernel 2.6.18-92.1.6.el5 the results are OK with the -r but not without (it failed at 1022 sockets without).

It makes me wonder though - does BIND (named) automatically do what the -r flag of selecttest is doing? Does it matter whether named is running as root or a different user? That is to say does the above test being successful with -r flag mean that I can be confident named is opening a wide range of ports?

-----Original Message-----
From: bind-users-bounce@isc.org [mailto:bind-users-bounce@isc.org] On Behalf Of JINMEI Tatuya / ????
Sent: Wednesday, July 30, 2008 8:36 PM
To: Shumon Huque
Cc: bind-users@isc.org
Subject: Re: selecttest tool

At Wed, 30 Jul 2008 19:52:02 -0400,
Shumon Huque wrote:

> I'd suggest adding code in the beginning to raise RLIMIT_NOFILE
> to the processes hardlimit, eg.:


Makes sense, especially because BIND9 does the same thing, and this
tool is primarily intended to check the behavior of BIND9. I've
updated the tool and put the revised version at the same URL:

http://www.jinmei.org/selecttest.tgz

---
JINMEI, Tatuya
Internet Systems Consortium, Inc.
----------------------------------
CONFIDENTIALITY NOTICE: This e-mail may contain privileged or confidential information and is for the sole use of the intended recipient(s). If you are not the intended recipient, any disclosure, copying, distribution, or use of the contents of this information is prohibited and may be unlawful. If you have received this electronic transmission in error, please reply immediately to the sender that you have received the message in error, and delete it. Thank you.
----------------------------------