using ksh or posix sh coprocessor |& with telnet on HP-UX 11 works but not on 11.11

[Belina Ang_1]

Does anyone have any idea why using the script below on HP-UX 11 works and not on HP-UX 11.11?

HOST=$1
ID=$2
PW=$3
telnet $HOST 2>&1 | tee /tmp/logfile |&
(sleep 1; print -p $ID)
(sleep 1; print -p $PW)
(sleep 1; print -p echo hello)
a="$(sleep 1; read -p | tr -d '\r')"
while [[ "$a" != *hello* ]] ;
do
print [$a]
a="$(sleep 1; read -p | tr -d '\r')"
done
print -p echo exit

Thanks in advance....

[Michael Steele_2]

Why don't you add debug lines? And see's how far it gets?

ksh: set -xv

sh: sh -x -v script

[Belina Ang_1]

Did that already as well as tusc, but was not able to see anything important except in the tusc file. On HP-UX 11.11, it seems to be using a select() instead of a sigtimedwait().

[Michael Steele_2]

You do have one of the following on the first line, right?

#!/usr/bin/ksh

#!/usr/bin/sh

[Ramkumar Devanathan]

hi,

I've seen this problem with 11.11 machines. But somehow now I am not able to reproduce this on the same hp-ux 11.11 machine. Check if this is a problem with patches required.

In any case, I suggest you use something like remsh to do the above. or even expect scripts.

http://hpux.connect.org.uk/hppd/hpux/Tcl/expect-5.37/

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x2919237a4bc6d611abdb0090277a778c,00.html

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x9ec691ccb36bd611abdb0090277a778c,00.html

- ramd.

[Belina Ang_1]

Yes, there is a
!#/usr/bin/ksh -x
on the first line. I am more inclined to think that patches would solve the problem. I just have not found the correct one yet. If anyone remembers, would really appreciate it.

[Jeff Schussele]

Hi Belina,

That first line HAS to be Sh-Bang - like

#!/usr/bin/ksh -x

NOT

!#/usr/bin/ksh -x

Was that just a typo in your last post?

Rgds,
Jeff

[Belina Ang_1]

Sorry, typo

it was a ....

#!/usr/bin/ksh

[Ramkumar Devanathan]

Belina,

I don't know if this will help very much, but my ksh binary has this what string -

# what /usr/bin/ksh
/usr/bin/ksh:
$Revision: B.11.11_LR
Wed Nov 8 19:55:23 PST 2000 $
$ B.11.11_LR Oct 27 2000 00:40:10 $
Version 11/16/88

I ran the following script several times just now -
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
#!/usr/bin/ksh

cat /var/adm/syslog/syslog.log |&
while read -p line
do
echo $line
done
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

And it worked fine.

Probably you may try to increase some kernel parameters - max_thread_proc is 1024 on this machine.

- ramd.

[Steven E. Protter]

First thing I'd do is check the PATH varable. All the commands you are using were migrated to 11.11 as far as I know. If its a new system, the /etc/profile might be different, causing different behavior.

SEP

[Belina Ang_1]

Hello Ramd,

=> what /usr/bin/ksh
/usr/bin/ksh:
$Revision: B.11.11_LR
Wed Nov 8 19:55:23 PST 2000 $
$ B.11.11_LR Oct 27 2000 00:40:10 $
Version 11/16/88

The script you wrote also worked on my 11.11

I have exactly the same results you got. I think I might be looking for a telnet patch since when I compared the tusc outputs between a 11 and 11.11. It started being different at the point when the password is about to be entered. I was ablt to enter the login name, but when it prompted for the password, I was getting the following:

Password: Local flow control on
Connection closed by foreign host.


19174: select(2048, 0x40005860, 0x40005760, 0x40005660, NULL) = 1
19174: recv(4, 0x400064f8, 1024, 0) = 0
19174: ioctl(0, TCSETATTRD, 0x7aff09c4) ERR#25 ENOTTY
19174: ioctl(0, TIOCSLTC, 0x40004b30) ERR#25 ENOTTY
19174: ioctl(0, FIONBIO, 0x7aff0a08) ERR#25 ENOTTY
19174: ioctl(1, FIONBIO, 0x7aff0a08) ERR#25 ENOTTY
19174: close(4) = 0
19174: ioctl(0, TCSETATTRD, 0x7aff0a04) ERR#25 ENOTTY
19174: ioctl(0, TIOCSLTC, 0x40004b38) ERR#25 ENOTTY
19174: ioctl(0, FIONBIO, 0x7aff0a48) ERR#25 ENOTTY
19174: ioctl(1, FIONBIO, 0x7aff0a48) ERR#25 ENOTTY
19174: select(2, NULL, 0x7aff09b8, NULL, NULL) = 1
19174: ioctl(0, TCSETATTRD, 0x7aff0a04) ERR#25 ENOTTY
19174: ioctl(0, TIOCSLTC, 0x40004b30) ERR#25 ENOTTY
19174: ioctl(0, FIONBIO, 0x7aff0a48) ERR#25 ENOTTY
19174: ioctl(1, FIONBIO, 0x7aff0a48) ERR#25 ENOTTY

[Ramkumar Devanathan]

Ok Belina,

I guess that explains it. Your connection got closed and so, the pipe was broken!

I think I might have faced a similar problem in reading a file which was probably not present. what are the errors you get when you run the script actually?

- ramd.

[Dietmar Konermann · Accepted solution]

I digged a little bit around this symptom.

The problem is not 11.11 specific. Instead it applies for 11.00 and 11.11, but was actually introduced by patches.

In fact the fix for JAGab75328 (11.00 PHNE_22159, 11.11 PHNE_23937) introduced it. With this fix the connection is supposed to be dropped if "stty 0" is issued. As a side effect telnetd closes connections when the speed is "0"... unfortunately this is true when telnet's stdin is no tty.

JAGae02032 adresses this incompatibility. It is included in the latest telnetd patches (11.00 PHNE_26096, 11.11 PHNE_24829) and restores the original behaviour, unless telnetd's new "-y" option is used.

Best regards...
Dietmar.

[Belina Ang_1]

Thank you everybody for your help.

=) Thanks again....