[hercules-390] hyperion cleaning

[[e-mail] [hercules-390]]

Hi Everyone,


I have a problem with hyperion:
If I start a session on a freshly booted Linux system everything is fine.
On the contrary starting anew a session of Hyperion when it was
properly exit leads to problems.
1) tun interface not properly closed.
2) tcp sessions with hosts 3270 busy.
3) socket from card reader busy.
4) 0E printer beheaving badly.
Those problems don't happen all the time.
Sometime only one of them but all the time something is wrong

among that list.


So you could say that the cleaning isn't properly done at the exit.
What I am doing wrong?


Thanks for your help.

[[e-mail] [hercules-390]]

We don't know, you haven't provided any information to give us a clue.

What actions do you take from the moment you decide to exit until the moment you believe "it was properly exit"?

[[e-mail] [hercules-390]]

Hi Mcisho,

Post by ***@yahoo.com [hercules-390]
What actions do you take

Well, I stop z/OS and after I wrote exit into the hercules console
and press enter.

[[e-mail] [hercules-390]]

If that's the best information you can provide then no-one is going to be able to tell what you're doing wrong.

[[e-mail] [hercules-390]]

Hi Mcisho,


What do you want to know and from where I have to start?


That has never happen with the same procedure Under hercules 3.12

[Joe Monk [e-mail] [hercules-390]]

Jean-Louis,

We have a practice on these lists that we sign our name to our emails.
Please adhere, otherwise you will be called no-name.

Joe

Post by ***@stben.net [hercules-390]
Hi Mcisho,
What do you want to know and from where I have to start?
That has never happen with the same procedure Under hercules 3.12

['\'Fish\' (David B. Trout)' [e-mail] [hercules-390]]

(pseudo-piggyback)

Post by ***@yahoo.com [hercules-390]
If that's the best information you can provide then no-one
is going to be able to tell what you're doing wrong.

Indeed.

This person who we can only address by their email address (***@stben.net) instead of by their name since they failed to identify who they are should probably take a moment to read the following helpful bit of advice:

https://github.com/Fish-Git/AskForHelp


Please forgive me who-ever-the-heck-you-are, but I'm not trying to pick on you. I'm not. I'm simply using your request for help as a good example of how NOT to ask for help in a public support forum. And I'm only doing this in order to help YOU. (And any others who might also be reading this thread)

If you would simply provide more detail you'd have a better chance of finding someone willing, and more importantly, *able*, to actually help you.

[Ivan Warren [e-mail] [hercules-390]]

Post by ***@stben.net [hercules-390]
Hi Everyone,
If I start a session on a freshly booted Linux system everything is fine.
On the contrary starting anew a session of Hyperion when it was
properly exit leads to problems.
1) tun interface not properly closed.

How do you determine that ?

Post by ***@stben.net [hercules-390]
2) tcp sessions with hosts 3270 busy.

How do you determine that ?

Post by ***@stben.net [hercules-390]
3) socket from card reader busy.

How do you determine that ?

Post by ***@stben.net [hercules-390]
4) 0E printer beheaving badly.

How do you determine that ?

Post by ***@stben.net [hercules-390]
Those problems don't happen all the time.
Sometime only one of them but all the time something is wrong
among that list.

--Ivan


[Non-text portions of this message have been removed]

[Laddie Hanus [e-mail] [hercules-390]]

--------------------------------------------
On Sun, 2/19/17, ***@stben.net [hercules-390] <hercules-***@yahoogroups.com> wrote:

Subject: [hercules-390] hyperion cleaning
To: hercules-***@yahoogroups.com
Date: Sunday, February 19, 2017, 7:43 AM


Hi
Everyone,
I have a problem with
hyperion:If I start a session on a freshly booted
Linux system everything is fine.On the contrary
starting anew a session of Hyperion when it was
properly exit leads to problems.1)
tun interface not properly closed.2) tcp sessions
with hosts 3270 busy.3) socket from card reader
busy.4) 0E printer beheaving
badly.Those problems don't happen all the
time.Sometime only one of them but all the time
something is wrong
among that
list.
So you could say
that the cleaning isn't properly done at the
exit.What I am doing wrong?
Thanks for your help.



Person with no name and cant include enough of the post they reply for others to understand what the reply is to


Did you close and and all 3270 windows
close any external socket programs
anything that uses the tun interface,
Basically anything external to hercules that talks to it.

Laddie Hanus

[Joe Monk [e-mail] [hercules-390]]

OK Guys. Enough.

His name is Jean-Louis. He has put it in previous emails.

Joe

Post by Laddie Hanus ***@yahoo.com [hercules-390]
--------------------------------------------
Subject: [hercules-390] hyperion cleaning
Date: Sunday, February 19, 2017, 7:43 AM
Hi
Everyone,
I have a problem with
hyperion:If I start a session on a freshly booted
Linux system everything is fine.On the contrary
starting anew a session of Hyperion when it was
properly exit leads to problems.1)
tun interface not properly closed.2) tcp sessions
with hosts 3270 busy.3) socket from card reader
busy.4) 0E printer beheaving
badly.Those problems don't happen all the
time.Sometime only one of them but all the time
something is wrong
among that
list.
So you could say
that the cleaning isn't properly done at the
exit.What I am doing wrong?
Thanks for your help.
Person with no name and cant include enough of the post they reply for
others to understand what the reply is to
Did you close and and all 3270 windows
close any external socket programs
anything that uses the tun interface,
Basically anything external to hercules that talks to it.
Laddie Hanus

[[e-mail] [hercules-390]]

Post by ***@stben.net [hercules-390]
What do you want to know and from where I have to start?

In an earlier message you said you "stop z/OS". The running jobs? JES? Quiesce?

You also said you "wrote exit into the hercules console and press enter". Had the stop z/OS actions completed when you pressed enter? What did you do after you pressed enter?

In summary, I want to see a log of everything you did and everything you saw when you shutdown Hercules.

How long after the first instance of Hercules was shutdown was the second instance of Hercules started? A second? A minute? An hour? And as Ivan asked, how did you determine there were problems?

This is the very minimum of information I, or anyone else, will need.

Ian

['Jean-Louis Noel' [e-mail] [hercules-390]]

Hi Ian,

Post by ***@yahoo.com [hercules-390]
In summary, I want to see a log of everything you did and
everything you saw when you shutdown Hercules.

Before everything, sorry to call you Mcisho.

Now, with z/OS :

- All batches finished.

- All remote connection closed but the console.

s shutall

*00.25.29 *02 BPXI078D STOP OF ZFS REQUESTED. REPLY 'Y' TO

* PROCEED. ANY OTHER REPLY WILL CANCEL THIS STOP.

r 2,y

00.26.05 $HASP099 ALL AVAILABLE FUNCTIONS COMPLETE

$p jes2

00.26.50 IEF404I JES2 - ENDED - TIME=00.26.50

z eod

00.27.29 IEE334I HALT EOD SUCCESSFUL

quiesce

Hercules:

As I said previously I typed ‘exit’ (without the inverted commas)

And everything was fine. I mean nothing unusual on the screen and the last line before the prompt is the same as the last line of the log that you can see at : http://www.stben.net/hercules/h2.log

The ‘next’ morning I restarted Hercules and it restarted with communication errors as you can see at: http://www.stben.net/hercules/h3.log

So, I stopped hercules with exit and that time it crashed with a segmentation fault trying to close a socket it never succeeded to open.



Another thing when it opens an interface, here tun0, with an IPV4 and an IPV6 address it add a route statement for the IPV4 but not for the IPV6.



Don’t bother to do anything for _ME_ because my copy of hercules is behaving properly now. The only purpose of this message is to state facts.

Bye,

Jean-Louis

[Ivan Warren [e-mail] [hercules-390]]

Jean Noel,

Please do a netstat -anp (from a root user) to see who is tying up port
3270.

--Ivan

Post by 'Jean-Louis Noel' ***@stben.net [hercules-390]
As I said previously I typed ‘exit’ (without the inverted commas)
And everything was fine. I mean nothing unusual on the screen and the
last line before the prompt is the same as the last line of the log
that you can see at : http://www.stben.net/hercules/h2.log
The ‘next’ morning I restarted Hercules and it restarted with
http://www.stben.net/hercules/h3.log

[Non-text portions of this message have been removed]

['Jean-Louis Noel' [e-mail] [hercules-390]]

Hi Ivan,

to see who is tying up port 3270

It could be you if you use a 3270 emulation terminal to connect to ns.stben.net:3270 but, use the IPV4 address because hyperon doesn’t respond in IPV6.



Now, being serious it’s the console task.



CNSLPORT 3270




#********************************************************************

# DISPLAY TERMINALS *

#********************************************************************

0700 3270 CONSOLE 10.149.85.0 255.255.255.0

0701 3270 MTSO 10.149.85.0 255.255.255.0

0702 3270 CICS 10.149.85.0 255.255.255.0

0704.13 3270



Bye,

Jean-Louis

[Ivan Warren [e-mail] [hercules-390]]

Post by 'Jean-Louis Noel' ***@stben.net [hercules-390]
Hi Ivan,

to see who is tying up port 3270

It could be you if you use a 3270 emulation terminal to connect to
ns.stben.net:3270 but, use the IPV4 address because hyperon doesn’t
respond in IPV6.
Now, being serious it’s the console task.

Jean Noel,

If I were to connect to port 3270 on your system, that would't tie it
up. It is tied up because hercules is attempting to bind[*] to local
address *:3270 but there is already another process bound to that
address. *:3270 means "IP_ADDR_ANY Port 3270" and usually used in
conjunction with a "listen()" to allow inbound connections to port 3270.

[*] The listen/server socket in the case of hercules is created/modified
with the SOCK_ADDR_REUSE flag that it is not prevented from binding even
if there are other sockets in the TIME_WAIT[**] state using that same port.
[**] a socket in TIME_WAIT state is a TCP session which has terminated.
However, because it is possible for some amount of time to still receive
extraneous datagrams, the TCP layer will keep a ghost version active
(usually for 30 seconds or so) so as to prevent the TCP stack from
flooding the network with "FIN" or "FIN/ACK" packets. However, without
ADDR_REUSE, the port is still going to be bound.

That's why we need the status of "netstat -anp" when you receive the
hercules message "HHC01023W Waiting for port 3270 to become free for
console connections"

If you have more than one instance running on your system, you need to
specify a different port for your inbound tn3270 connection on your
other instance. Note that you can specify any port that is not in use in
the range 1024-65535 if not running as root (ports 0-1023 are reserved
for privileged use).

--Ivan


[Non-text portions of this message have been removed]

['Jean-Louis Noel' [e-mail] [hercules-390]]

Hi Ivan,

Post by Ivan Warren ***@vmfacility.fr [hercules-390]
That's why we need the status of "netstat -anp" when you receive
the hercules message "HHC01023W Waiting for port 3270 to
become free for console connections",___

Only hercules uses the port 3270 on that machine.

And, as I stated on the first mail, it won’t happen again, I hope so, because I think I fixed it.



Bye,

Jean-Louis

[Ivan Warren [e-mail] [hercules-390]]

Post by 'Jean-Louis Noel' ***@stben.net [hercules-390]
Hi Ivan,

Post by Ivan Warren ***@vmfacility.fr [hercules-390]
That's why we need the status of "netstat -anp" when you receive
the hercules message "HHC01023W Waiting for port 3270 to
become free for console connections",___

Only hercules uses the port 3270 on that machine.
And, as I stated on the first mail, it won’t happen again, I hope so,
because I think I fixed it.

Jean Noel,

What was the fix ? Any issue with hercules ? Any issue terminating
hercules and hercules freeing the port ? (the port should be
automatically freed once all threads in the process have terminated).

--Ivan


[Non-text portions of this message have been removed]

['Jean-Louis Noel' [e-mail] [hercules-390]]

Hi Ivan,

Post by Ivan Warren ***@vmfacility.fr [hercules-390]
Please do a netstat -anp

Pay attention that the logs were from the 20th!

There you are: http://www.stben.net/hercules/netstat.txt



Bye,

Jean-Louis