inetd's (TCP/IP v5.0) strange -SIGHUP handling.

What do I do wrong?

If I add a service to the /etc/inetd.conf (for example I un-comment the
FTP entry there) and send “kill -SIGHUP [inetd’s PID]” - the super-server
survives and adds a service indeed.
If I try removing a service by commenting it’s entry out and send the same
“kill -SIGHUP [inetd’s PID]” - the super-server allways dies…

This does not depend on the actual service added|removed, inetd allways
adds successfully and allways fails to survive the removal.

QNX v4.25G + Proc32 v4.25Q, TCP/IP v5.0

Tony.

On Wed, 03 Aug 2005 02:43:41 +0400, Tony <mts.spb.suxx@mail.ru> wrote:

QNX v4.25G + Proc32 v4.25Q, TCP/IP v5.0
The fact of Proc32 being the v4.25Q seems to be irrelevant, I’ve rolled

back to Proc32 v4.25O and it is the same behavior for /usr/ucb/inetd

Tony.

Tony <mts.spb.suxx@mail.ru> wrote:

What do I do wrong?

I don’t think you did anything wrong.

If I add a service to the /etc/inetd.conf (for example I un-comment the
FTP entry there) and send “kill -SIGHUP [inetd’s PID]” - the super-server
survives and adds a service indeed.
If I try removing a service by commenting it’s entry out and send the same
“kill -SIGHUP [inetd’s PID]” - the super-server allways dies…

Yup, does it for me, too.

Looks like a bug in inetd to me.

I’ve issued a PR against this, but do not have any idea when or if a
fix might be made.

I think that:
slay inetd; inetd

would work as a reasonable work around.
(Or,if you prefer:
kill [inetd’s pid]; inetd
But, I like slay, as it saves me having to figure out inetd’s pid from
the ps or sin listing.)

-David

David Gibbs
QNX Training Services
dagibbs@qnx.com

On Wed, 03 Aug 2005 19:44:48 +0400, David Gibbs <dagibbs@qnx.com> wrote:

Tony <> mts.spb.suxx@mail.ru> > wrote:
What do I do wrong?
I don’t think you did anything wrong.
:frowning:



“kill -SIGHUP [inetd’s PID]” - the super-server allways dies…
Yup, does it for me, too.
Looks like a bug in inetd to me.
:frowning:



I think that:
slay inetd; inetd
would work as a reasonable work around.
Yes, but it changes to the newer program group if started from the shell

session…
Does not look that cute in “sin se” report.

Well, I hope XTang will have a minute to look into this.

Tony.