Hi,

we are also seeing the same sympthoms since upgrading to NB 4.5FP3-5.
We have 2 Master Servers in 2 different domains, and each one shows the same
sympthoms.
Masters are SUN Ultra2 with 2GB(!) of RAM and Solaris 7 and 9.

After having started the java Console for about 2 days, all the memory on
the server is used up.
It looks like a memory leak in the java client. You can see with the top
commando, how the memory is slowly counting down.

==> Hope Veritas will fix this soon, this Bug is very dangerous, sometimes
then i forget to close the GUI, memory is used up and Backup prozesses die!


Regards,
Klaus Micheler

-----Original Message-----
From: veritas-bu-admin AT mailman.eng.auburn DOT edu
[mailto:veritas-bu-admin AT mailman.eng.auburn DOT edu] On Behalf Of Cyril
Gnanaprakasham
Sent: Mittwoch, 29. Oktober 2003 07:37
To: 'David Rock'; veritas-bu AT mailman.eng.auburn DOT edu
Subject: RE: [Veritas-bu] NBU 4.5 FP5 Activity Monitor Allocation Failed
Error


Well Iam facing the same problem. We did an upgrade from 3.41 to 4.5 and now
my activity monitors throws the same error. Any solutions on this 

Cyril

-----Original Message-----
From: veritas-bu-admin AT mailman.eng.auburn DOT edu
[mailto:veritas-bu-admin AT mailman.eng.auburn DOT edu] On Behalf Of David Rock
Sent: Wednesday, October 29, 2003 11:52 AM
To: veritas-bu AT mailman.eng.auburn DOT edu
Subject: Re: [Veritas-bu] NBU 4.5 FP5 Activity Monitor Allocation Failed
Error


* David Rock <dave-bu AT graniteweb DOT com> [2003-10-28 13:35]:
> We recently upgraded to 4.5 FP 5 from 3.4.3 and when using the java
> client GUI, we are getting a situation where the first couple days are

> great, but then we start seeing an error message pop up:
> 
>       Activity Monitor failed and must be restarted. To recover:
>       
>       1) Exit Activity Monitor
>       2) Correct the problem that caused the error
>       3) Restart Activity Monitor
> 
>       allocation failed
> 
> 
> The problem is, this is happening on ALL java client connections to
> the master server, Linux, Sun, Windows, etc. We have also noticed that

> over the last week since the upgrade, there are a LOT of java
> processes that seem to be old processes from client connections that 
> no longer exist that show up on the master server.
> 

Well, I am posting to myself, but it is pertinant information...

We have discovered that the problem IS resources being takeen up on the
master server, and the reason is because of stale connections to
clients. Most of them are Windows-based java clients that have been
ungracefully exited (powered off machine, GUI lockup, etc). We made sure
all client GUI's were logged off and killed all the remaining processes
on the master and life is good again. 

Anyone know how to get the master to clean up stale connections if a
client dies?

-- 
David Rock
david AT graniteweb DOT com

_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu