SQL Anywhere database - ASA 8.03 Netware 5

Author

ASA 8.03 Netware 5

snelbert

2005-08-19, 1:24 pm

I have ASA 8.03 installed on a Netware 5.0 server. An .ncf file loads the
Dbsrv8.nlm as shown here:

LOAD sys:asany\dbsrv8 -ct- -c 25M -tq 22:00 -ti 0 -tl 0 -x
TCPIP{myip=192.168.250.1;DoBroadcast=No}
vol1:access/data/accessdb8/accessdb.db

CRON is used to load run the ncf file each morning. Sometimes the nlm does
not load. It part loads and then fails with the error message "unable to
start on default port", it then goes on to say it will try port 49152
instead. The only way I can resolve this issue is to reboot the server. I
have netware 6 servers running the same ncf and there are no problems
reported. This happens on both Netware 5 servers.

Also you will notice that -tq unloads the database. If a client is still
connected to the database the nlm does not unload, but does not let other
clients connect thereafter. The (3rd party) backup then fails to backup the
database file because it is still held presumably, by the database server.
Again the only way to resolve this is to reboot the server and the client.

Has anyone anything like any of these issues before?

Snelbert

Nick Elson

2005-08-24, 11:24 am

Try removing the -ti 0 and -tl 0 switches.
Those are probably causing some of your problems
[especially for the -tq]

If you must have large timeouts, set those to non-zero
but bigger values than the defaults. But with larger (or
the infinite, with 0) timeouts you will likely be causing
TCP/IP connection to persists longer and that can
cause sockets to not get cleaned up.

There could be a Netware Service Patch that affects
this but you would need to talk to Novell about that
possibility.

On a related note, there is a significant change in the
ASA 9 architecture for Netware. Going there could
also affect this different. Just be aware ASA 9 will
require 5.1 SP6 (or possible something higher)
or 6.0SP 3 or 6.5 to run ASA 9.

Note that the use of alternate ports is an 'agility' feature
of our server that clients know how to navigate and
connect using this. If your clients are failing to track
this that could be due to Netware 5's stack preventing
UDP packets from discovering this server. A workaround
to this is to add multiple ports to the client's connection
strings

& #91;comm]LINKS=tcpip
& #123;port=2638,49152
,....}

good luck

"snelbert" < snelbert@REMOVEtotal
ise.co.uk> wrote in message
news:4306092c@forums
-1-dub...
>I have ASA 8.03 installed on a Netware 5.0 server. An .ncf file loads the
> Dbsrv8.nlm as shown here:
>
> LOAD sys:asany\dbsrv8 -ct- -c 25M -tq 22:00 -ti 0 -tl 0 -x
> TCPIP{myip=192.168.250.1;DoBroadcast=No}
> vol1:access/data/accessdb8/accessdb.db
>
> CRON is used to load run the ncf file each morning. Sometimes the nlm
> does
> not load. It part loads and then fails with the error message "unable to
> start on default port", it then goes on to say it will try port 49152
> instead. The only way I can resolve this issue is to reboot the server.
> I
> have netware 6 servers running the same ncf and there are no problems
> reported. This happens on both Netware 5 servers.
>
> Also you will notice that -tq unloads the database. If a client is still
> connected to the database the nlm does not unload, but does not let other
> clients connect thereafter. The (3rd party) backup then fails to backup
> the
> database file because it is still held presumably, by the database server.
> Again the only way to resolve this is to reboot the server and the client.
>
> Has anyone anything like any of these issues before?
>
> Snelbert
>
>
>

Sponsored Links