Problem with event channels

[Kretschel Klaus]

Hi,

I try to run 2 instances of a MICO application in the same network; the
applications communicate via event channels. To separate the 2 instances I
use different port numbers that are provided to micod via the
options "-ORBIIOPAddr inet:hostname:port --forward"
and to eventd via the options "-ORBImplRepoAddr
inet:hostname:port -ORBNamingAddr inet:hostname:port". nsd is started with
the command "imr create NameService poa
nsd "IDL:omg.org/CosNaming/NamingContext:1.0#NameService" -ORBImplRepoAddr
inet:hostname:port".

In the moment the second instance starts its micod some processes of the first
instance receive a message for the second time that was issued at startup
time of application 1 (and that is not valid in the current state).

How is it possible that startup of micod with a different port number disturbs
the other application?

Is it possible to avoid this problem?

I run MICO 2.3.11 under Suse Linux 9.2 with gcc 3.3.3.

Thanks for any information,
Klaus

PS: If I remember correctly this effect did not occur when I started the
applications under different accounts; her I used the same account.

[Karel Gardas]

Hi,

just short question: are you sure your second application imr does not
work against the first application micod? If so, then please try to
provide more information, even better might be simple example including
all starting scripts which duplicates this behavior.

Cheers,
Karel

Post by Kretschel Klaus
Hi,
I try to run 2 instances of a MICO application in the same network; the
applications communicate via event channels. To separate the 2 instances I
use different port numbers that are provided to micod via the
options "-ORBIIOPAddr inet:hostname:port --forward"
and to eventd via the options "-ORBImplRepoAddr
inet:hostname:port -ORBNamingAddr inet:hostname:port". nsd is started with
the command "imr create NameService poa
nsd "IDL:omg.org/CosNaming/NamingContext:1.0#NameService" -ORBImplRepoAddr
inet:hostname:port".
In the moment the second instance starts its micod some processes of the first
instance receive a message for the second time that was issued at startup
time of application 1 (and that is not valid in the current state).
How is it possible that startup of micod with a different port number disturbs
the other application?
Is it possible to avoid this problem?
I run MICO 2.3.11 under Suse Linux 9.2 with gcc 3.3.3.
Thanks for any information,
Klaus
PS: If I remember correctly this effect did not occur when I started the
applications under different accounts; her I used the same account.
_______________________________________________
Mico-devel mailing list
http://www.mico.org/mailman/listinfo/mico-devel

[Kretschel Klaus]

Hi,

Post by Karel Gardas
Hi,
just short question: are you sure your second application imr does not
work against the first application micod? If so, then please try to
provide more information, even better might be simple example including
all starting scripts which duplicates this behavior.

No, I'm not sure since I did not really understand the imr arguments (and also
Arno Puder's etc. book that I just got a few days ago did not help much on
this topic).

The software is much too big to supply it, but the start-up scripts (condensed
to the essential parts) may hopefully supply enough information:

# Usage: start_mico.sh MICO_ADDR (i.e. port number)

MICO_ADDR=$1
RC="-ORBImplRepoAddr $1 -ORBNamingAddr $1"

# Run BOA daemon (micod)
micod -ORBIIOPAddr ${MICO_ADDR} --forward &

# Register name service (nsd)
imr create NameService poa \
nsd "IDL:omg.org/CosNaming/NamingContext:1.0#NameService" \
-ORBImplRepoAddr ${MICO_ADDR}

# Run event service (eventd)
eventd $RC &

BTW:
As a work-around we use different accounts and our software is running. But an
answer would of course be fine for my understanding of what happens here.

Thank you,
Klaus

Post by Karel Gardas
Cheers,
Karel

Post by Kretschel Klaus
Hi,
I try to run 2 instances of a MICO application in the same network; the
applications communicate via event channels. To separate the 2 instances
I use different port numbers that are provided to micod via the
options "-ORBIIOPAddr inet:hostname:port --forward"
and to eventd via the options "-ORBImplRepoAddr
inet:hostname:port -ORBNamingAddr inet:hostname:port". nsd is started
with the command "imr create NameService poa
nsd "IDL:omg.org/CosNaming/NamingContext:1.0#NameService"
-ORBImplRepoAddr inet:hostname:port".
In the moment the second instance starts its micod some processes of the
first instance receive a message for the second time that was issued at
startup time of application 1 (and that is not valid in the current
state).
How is it possible that startup of micod with a different port number
disturbs the other application?
Is it possible to avoid this problem?
I run MICO 2.3.11 under Suse Linux 9.2 with gcc 3.3.3.
Thanks for any information,
Klaus
PS: If I remember correctly this effect did not occur when I started the
applications under different accounts; her I used the same account.
_______________________________________________
Mico-devel mailing list
http://www.mico.org/mailman/listinfo/mico-devel