MAXACCESSERS and free BBL slots

Paul Seager-Smith

unread,

Aug 11, 1999, 3:00:00 AM8/11/99

to

We have a an application which is acting as a Tuxedo client via domain
gateways to a Tuxedo Server. On the client side we have approx. 10
client processes. However, we have sometimes been getting messages on
the client machine that
the BBL is full:

091336.triumph!?proc.29386: LIBTUX_CAT:1122: ERROR: No space in Bulletin
Board
091336.triumph!?proc.29386: LIBTUX_CAT:582: ERROR: Unable to register,
registry
table full

Since the BB defaults to setting MAXACCESSERS to 50 (tmconfig confirms
this), we are puzzled as
to why the BB should be overflowing with only the 10 clients that we
have. The TMIB is also only showing up the clients that are connected
(although access to the TMIB is also difficult when this problem
occurs!)

Any explanation would be most welcome.

Rgds,

Paul Seager-Smith

Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.

Michel Roman

unread,

Aug 11, 1999, 3:00:00 AM8/11/99

to

Paul,

BB slots are use for client and servers (administrative, user). Your
number of servers may be high enough so that you don't have enough BB
slot left for client processes. You may want to consider increasing the
MAXACCESSERS and the MAXSERVERS parameters of the *RESOURCE section to
solve your problem.

Michel

BEA Systems, Inc.
www.beasys.com
__________________________________________________________

The views expressed in this posting are solely those of the author,
and BEA Systems, Inc. does not endorse any of these views. BEA
Systems, Inc. is not responsible for the accuracy or completeness of
the information provided and assumes no duty to correct, expand upon,
delete or update any of the information contained in this posting.
___________________________________________________________

Scott Orshan

unread,

Aug 11, 1999, 3:00:00 AM8/11/99

to

Every TUXEDO process counts as an accesser, including the Domain
gateways, the WSH/WSL, and TUXEDO servers you are running, TMSs, etc. If
you have MAXWSCLIENTS set, it takes away from MAXACCESSERS.

If you don't think you have exceeded MAXACCESSERS, check whether you
have clients that exit without calling tpterm. These would continue to
use resources until they are cleaned up by the BBL.

Scott Orshan
BEA Systems

Paul Seager-Smith wrote:
>
> We have a an application which is acting as a Tuxedo client via domain
> gateways to a Tuxedo Server. On the client side we have approx. 10
> client processes. However, we have sometimes been getting messages on
> the client machine that
> the BBL is full:
>
> 091336.triumph!?proc.29386: LIBTUX_CAT:1122: ERROR: No space in Bulletin
> Board
> 091336.triumph!?proc.29386: LIBTUX_CAT:582: ERROR: Unable to register,
> registry
> table full
>
> Since the BB defaults to setting MAXACCESSERS to 50 (tmconfig confirms
> this), we are puzzled as
> to why the BB should be overflowing with only the 10 clients that we
> have. The TMIB is also only showing up the clients that are connected
> (although access to the TMIB is also difficult when this problem
> occurs!)
>
> Any explanation would be most welcome.
>
> Rgds,
>
> Paul Seager-Smith
>
> Sent via Deja.com http://www.deja.com/
> Share what you know. Learn what you don't.

BEA Systems, Inc.

Paul Seager-Smith

unread,

Aug 12, 1999, 3:00:00 AM8/12/99

to

You were right here, a value of MAXWSCLIENTS=40 had somehow slipped in
here, which makes the lack of BBL slots much easier to explain! As an
aside, how does the BBL notice that clients (particularly
non-conversational ones) have died and how often does it check ??

Many thanks.

Paul Seager-Smith

In article <7osp8g$j7d$1...@news.beasys.com>,

Scott Orshan

unread,

Aug 13, 1999, 3:00:00 AM8/13/99

to

Paul Seager-Smith wrote:
>
> You were right here, a value of MAXWSCLIENTS=40 had somehow slipped in
> here, which makes the lack of BBL slots much easier to explain! As an
> aside, how does the BBL notice that clients (particularly
> non-conversational ones) have died and how often does it check ??
>
> Many thanks.
>
> Paul Seager-Smith
>

Lucky guess!

Native clients are checked every SCANUNIT*SANITYSCAN seconds. Cleaning
up after dead processes is what distinguishes our software from lesser
grade products, so I don't want to give away our algorithms, but there
aren't too many ways to check for a dead process on UNIX (NT is
different), so your first guess would probably be correct.

WS clients are cleaned up when their network connection drops, or you
can use the Idle Client timer on the WSL command line to detect a dead
PC.

Scott Orshan

Michael R. Milani

unread,

Aug 13, 1999, 3:00:00 AM8/13/99

to

Read up on the SANITYSCAN, SCANUNIT parameter and the RESTART= ,GRACE=, and
MAXGEN= parameter.
These help set up the restart option.

Good Luck,
Michael R. Milani
michael...@walgreens.com

Paul Seager-Smith <seage...@my-deja.com> wrote in message
news:7ovblc$cjl$1...@news.beasys.com...

>
> You were right here, a value of MAXWSCLIENTS=40 had somehow slipped in
> here, which makes the lack of BBL slots much easier to explain! As an
> aside, how does the BBL notice that clients (particularly
> non-conversational ones) have died and how often does it check ??
>
> Many thanks.
>
> Paul Seager-Smith
>
>

> > BEA Systems, Inc.
> > www.beasys.com
> > __________________________________________________________
> >
> > The views expressed in this posting are solely those of the author,
> > and BEA Systems, Inc. does not endorse any of these views. BEA
> > Systems, Inc. is not responsible for the accuracy or completeness of
> > the information provided and assumes no duty to correct, expand upon,
> > delete or update any of the information contained in this posting.
> > ___________________________________________________________
> >
> >
>
>