Hi , i will give this week all the ideas i got a try, and i will inform you of the outcoming.

Thanks for the help up to now.

Hoffmann Marcel

From: Brent Frère [mailto:brent@bfrere.net]
Sent: dimanche 9 avril 2006 23:04
To: HOFFMANN Marcel
Subject: Re: [Lilux-help] Kill stops too much processes

The father process receives a signal when one of its sons dies. The default behaviour of a Unix process in this case is 'ignore', but it might be set a different way, and the default behaviour of a signal that is not ignored but unhandeled is 'kill' the process receiving the signal, so the father process.

Other sons might not survive this situation, especially if they are not completely independant from the father. As example, if they share pipes, and the last writer or reader dies, they might consider this as an inacceptable condition and stop.

I would also check the possible alias for 'kill' to make sure it was not aliased a way or an another.
Lets try
> alias kill
or use \kill to avoid a possible alias.

Maybe the processes showns are actually not different processes, but different threads in the same processus space. Can you confirm ?
I know also that there are 'light' forks on modern unices. They actually share the same memory area, and separate the memory pages only in case of writing. This is not far from the thread situation. Maybe it is the case here.

I think you should send an another signal than -9 in order to stop this process. -9 is quite hard, as the process has no chance to catch it: it is no more executed at all. An another signal, unhandeled, as -HUP, gives the process the possibility to close a cleaner way.

Tell me if it helps.

HOFFMANN Marcel a écrit:

Hi , to all Lilux Users,

We have migrated some of our AIX and Solaris Programs to Linux with some success. We are
hitting only smaller "features" like this one where we find no real explanations , even by googling.

So the problem is when we kill one process all the processes of the user are killed. But to make it
more complicated not all the time , so it is not reproductible.

If someone knows a hint to this ….

Thanks in advance

Hoffmann Marcel

**************************
[operator@pportlc1 monitor]$ ps -fu operator

UID        PID PPID C STIME TTY          TIME CMD
operator 14975 14974 0 10:26 ?        00:00:00 sh
operator 14976 14975 0 10:26 ?        00:00:00 /bin/bash
operator 14977 14976 0 10:26 ?        00:00:00 /bin/ksh ./lanceMonitor.ksh
operator 14978 14977 0 10:26 ?        00:00:00 /bin/ksh ./processMonitor.ksh /home/operator/echg/monitor/config/processmonitor.properties

operator 14979 14978 0 10:26 ?        00:01:51 /home/softs/Java/j2sdk1.4.1_01/jre/bin/java -cp lib/clientutil.jar:lib/mail.jar:lib/log4j-1.2.11.jar:lib/client

operator 23738 23737 0 11:50 ?        00:00:00 sh
operator 23739 23738 0 11:50 ?        00:00:00 /bin/bash
operator 23740 23739 0 11:50 ?        00:00:00 /bin/ksh ./Lance_VerifEchgGedb.ksh
operator 23741 23740 0 11:50 ?        00:00:01 /bin/ksh /home/operator/echg/bdl/bin/VerifEchgGedb.ksh operator /home/operator/echg/bdl/bin

operator 23745 23741 0 11:50 ?        00:00:00 /bin/ksh /home/operator/echg/bdl/bin/LanceEchgGedb.ksh
operator 23747 23745 1 11:50 ?        00:03:27 /home/softs/Java/j2sdk1.4.1_01/jre/bin/java -Dfile.encoding=iso-8859-1 -DHTTPClient.log.mask=-1 -cp /home/bea/w

operator 23822 23821 0 11:51 ?        00:00:00 sh
operator 23823 23822 0 11:51 ?        00:00:00 /bin/bash
operator 23824 23823 0 11:51 ?        00:00:00 /bin/bash
operator 23825 23824 0 11:51 ?        00:02:07 /home/softs/Java/j2sdk1.4.1_01/jre/bin/java -Dapiopc.repbase=/home/operator -Dapiopc.replog=/home/log/operator/

operator 24213 24201 0 17:05 pts/1    00:00:00 -bash
operator 24617 23741 0 17:09 ?        00:00:00 sleep 30
operator 24620 24213 0 17:09 pts/1    00:00:00 ps -fu operator

[operator@pportlc1 monitor]$ kill -9 14978

[operator@pportlc1 monitor]$ ps -fu operator

UID        PID PPID C STIME TTY          TIME CMD
operator 24213 24201 0 17:05 pts/1    00:00:00 -bash
operator 24661 24213 0 17:09 pts/1    00:00:00 ps -fu operator

********************************

**********************************************************************************************************
NOTICE
Electronic messages are not secure, error free, and can contain viruses or may be delayed; the sender is not liable for any of these occurences.
The information contained in this electronic message and any attachments (the "Message") is intended for one or more specific individuals or entities, and may be
confidential, proprietary, privileged or otherwise protected by law. Access to or use of this Message by anyone else is unauthorized. If you have received this Message in
error, please notify the sender immediately and delete it from your system. Any unauthorised use, dissemination, forwarding, printing or copying of this message is
prohibited.
Any views expressed in this Message are solely those of the individual sender, except where the sender specifically states them to be the views of the Luxembourg Stock
Exchange, or CCLux, or any other subsidiary of the Luxembourg Stock Exchange.
End of disclaimer
**********************************************************************************************************
_______________________________________________
Lilux-help mailing list
Lilux-help@lilux.lu
http://lilux.lu/mailman/listinfo/lilux-help
  

-- 
Brent Frère

Private e-mail:  Brent@BFrere.net

Postal address: 58, rue d'Esch
                L-3720 Rumelange
                Grand-Duchy of Luxembourg
                European Union

Mobile: +352-021/29.05.98
Fax:    +352-26.30.05.96
Home:   +352-307.341
URL:    http://BFrere.net

If you have problem with my digital signature, please install the appropriate authority certificate by browsing https://www.cacert.org/certs/root.crt.