Hi Gurus,
The methods suggested by all who responded are useful. I will try
implementing them as it applies.
Below are the list of responses that i received,
Andrew Ho wrote:
What you have described seems very strange. At my site, we have
hundreds of Nscp Enterprise servers running and we have not seem many
cases where httpd does not get restarted with restart, or stop then
start. I think you should get a core of the system when it gets into a
state like you have described and open a ticket with Netscape and Sun to
do some analysis. But before you do so, make sure you have all the new
solaris patches and Also what does the load look like on you machine
when this happens, there are times when the load becomes very high,
operating systems does not respond normally.
" The system performance is good and the loads are normal , I will be
sending a core to sun/netscape but it will end-up with they blaming each
other. Will give it a try ;-) thou "
Geoff Weller wrote:
This may help: will a kill -HUP work?
" I have already tried , not much success, Thanks for trying "
Jim Harmon wrote :
In your /etc/rc2.d dir you should have a process script called
"Snnhttpd" where "nn" is a number from 01 to 99, probably around 96 or
97.
You should be able to stop your httpd process(es) completely by doing
the following:
cd /etc/rc2.d
./Snnhttpd stop
Wait about 15 minutes, then use
./Snnhttpd start
to restart the deamon(s).
Chances are that the httpd deamon you are trying to kill with -9 is a
child process of a more "resilient" deamon that has a "keep alive"
feature running on the httpd process(es).
Using the rc script to stop the deamon should work where "kill" doesn't.
" I have made the attempt to stop the http process thru rc*.d scripts,
/etc/init.d scripts and the process still won't die, only then i went
ahead for a manual kill of all process that were running for http - not
much sucess -- appreciate your response "
K.Ravi wrote:
kill -9' won't work if a process is waiting for some I/O - like reading
from disk. Maybe that is the reason your web server hangs. Any disk
errors in /var/adm/messages?
" No disk errors seen, but i am sure the process is waiting for
something to exit, what could it be? i am still looking "
Tom Bosick wrote:
I have also encountered a similar problem. I installed the latest patch
from Netscape, which definitely reduced the problem. However, I still
get the problem. Please let me know what you find out
" I hope this summary helps, if you can send me some details on the
patch , that would be great "
Seth Rothenberg wrote:
I read somewhere (recently, maybe on this list?)
that you can send a SIGSTOP to defunct processes.
The kill man page says you can type kill -l
to get a list. From it, I see that SIGSTOP is 23.
I hope this helps.
" I taught defuncts needed a reboot, i will try it out - thank you "
Harvey Wamboldt wrote:
What's probably happening is that this is a child process that exits
and is waiting to be "reaped" by it's parent process, which can't
because it's waiting for something. These are called zombie
processes. They actually don't take up much space and unless you have
a lot of them, or you have a machine you can't reboot, they shouldn't
cause any problems. You can read about it in signal(3C), exit(2),
proc(4), sigaction(2), ps(1) and wait(2).
You might be able to kill the child process by killing it's parent.
There are also some occasions where it looks as though your attempt to
kill a child process failed, but what happened is that the parent
process simply forked off another child...
" Without killing the CHILD process my netscape server won't restart
and the http had no parent process forking childs as the process looked
like a defunct with PID 1 "
Thank you all for the responses,
raj
--------------
My Original Posting ,
Hi Gurus,
>
> I have been a member of this list for a few months now and i am >
amazed at the amount of resources i get from you all out there.
> I maintain Sun's with 2.4, 2.5 , 2.6 running on 4u, 4m arch's , i >
am also running Nestcape Web Server 2.x/3.x on this systems , >
recently i have come across a problem in that at times when the web >
server stops responding a restart does not help and i need to >
manually kill the http process . But strangely at times the http >
process refuses to die with a kill or a kill -9 ,
>
> truss on kill -9 "Pid" does not help much.
>
> The only option left for me is to bring the http server backup is to
> reboot the system which i have to do each time this happens. I was
> hoping to know if someone has come across this ans is there any way >
i can get rid of this process other than a reboot.
>
> I have skimmed the archives but any response would be greatly
> appreciated, and I will summarise.
>
> Thank you
>
> raj
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:36 CDT