[SunHELP] Activeworks!!!

Thu Mar 29 22:59:43 CST 2001

Sir:

DBAs generally don't know operating systems; your DBA
is no different.  TIME_WAIT and CLOSE_WAIT are normal
states for a socket.  They occur while a socket is closing.
There *can* be problems where a socket gets stuck in
either state, which requires a reboot to "fix."  However,
the fact that the sockets remain in either state doesn't
mean that there's something *debilitatingly* wrong with
the "TCP/IP stack"; it simply means that the sockets
are stuck in that state.  It's likely that this is a
busy system: sometimes the sockets will hang around
in that state for quite a while before all of the
appropriate TCP packets are passed between the client
and the server.  Granted, the clients of all the
sockets you show in the netstat output are on the 
same machine as the server.  I just want you first
to understand that while sockets can get stuck in
those states, that fact alone doesn't mean that
non-closed sockets (or even new ones) can't work
just fine.

All you need to do to clear this up with the DBA
is show him or her that you can telnet to port 1521
without getting the message "Connection refused."
If you can, there's no problem.  If you can *and*
the sockets still stick around in CLOSE_WAIT *forever*
or until a reboot, that's a separate problem, and
you don't have to suffer the pressure of the DBAs
(likely) application-related problem while attempting
to solve it.  Rather, you can take your own sweet
time researching it.

If you telnet to port 1521 and get a "Connection 
refused" error, first check whether the application
is indeed listening on that socket.  You should see
a socket with a state of LISTEN in the output of
netstat.  If you don't see a socket in the LISTEN
state, the application hasn't done a bind(2) and
listen(2).  (Those are operating system calls.)
If you *do* see a socket on that port in the LISTEN
state, and you still get a "Connection refused" error,
then you definitely have an operating system-related
problem.

This, by the way, *might* be yet another classic
scenario in the world of computers: 

	* Application administrator has problem.
	* Application administrator is not sufficiently
	  talented or trained to solve the problem, or
	  perhaps fails to read documentation.
	* Application administrator blames operating
	  system, and brings hail down on the head
	  of the sysadmin.
	* Sysadmin doesn't know enough to fight back.
	* Sysadmin suffers.

I suggest, finally, that you find a good book 
on TCP/IP with which to defend yourself.  (The
ones by Stevens are good.  I'd have had more
precise information on CLOSE_WAIT and TIME_WAIT
were my books by him not packed in boxes at
the moment.)

Good luck.

Nicholas Dronen

On Thu, Mar 29, 2001 at 10:02:27AM -0000, ES Srinivas wrote:
> Hi All
>    We have a Sun E250 running on Solaris 7(SunOS sun250b 5.7 Generic_106541-11 sun4u sparc SUNW,Ultra-250),which is also having Oracle,Activeworks,JDE Oneworld Adapter installed.
>    The problem is the application Activeworks is used to write some log files into the Oracle database,which are both on the same system.At the application end the following is the error:
> 
> 11:43:02 AM [Debug] Logger(0)                                SQLException in update restart count: Error Message: Exhausted Resultset; Error Code: 17011; SQLState: null
> 11:43:02 AM [Debug] Integration(2)                           Unable to log begin event for com.wipro.security.active.atc.iscatc.ISCAgentForWs; Exception: SQLException in update restart count: Error Message: Exhausted Resultset; Error Code: 17011; SQLState: null
> 11:43:02 AM [Adapter] Unable to log begin event, is logging database available?
> 11:43:02 AM Unable to log begin event, is logging database available?
> 
>    The DBA says there is nothing wrong with Oracle and found out that there is a problem at the TCP/IP socket level with the following output of netstat command.
>  
> 10.150.50.112.35032  10.150.50.112.1521   32768      0 32768      0 CLOSE_WAIT
> 10.150.50.112.35040  10.150.50.112.34902  32768      0 32768      0 CLOSE_WAIT
> 10.150.50.112.34369  10.150.50.116.1521    8760      0  8760      0 TIME_WAIT  
> 10.150.50.112.34676  10.150.50.118.59305   8760      0  8760      0 CLOSE_WAIT
> 10.150.50.112.35025  10.150.50.112.1521   32768      0 32768      0 CLOSE_WAIT
> 10.150.50.112.35041  10.150.50.112.33327  32768      0 32768      0 CLOSE_WAIT
> 10.150.50.112.35042  10.150.50.112.33605  32768      0 32768      0 CLOSE_WAIT
> 
>    According to the DBA 1521 is the port on which Oracle listens.
>    The user says Active Works is  a dynamic port configurer and the DBA says its only the problem with TCP/IP stack or socket level of Solaris and is for the System Admin to resolve the issue.
>     Can somebody help me 
>     This is very Urgent
>     Thanks in Advance
> 
> Srinivas.E.S.
> 
> 
> _____________________________________________________
> Chat with your friends as soon as they come online. Get Rediff Bol at
> http://bol.rediff.com
> 
> 
> 
> 
> _______________________________________________
> SunHELP maillist  -  SunHELP at sunhelp.org
> http://www.sunhelp.org/mailman/listinfo/sunhelp