UNIX Auto Shut Down and Start Up of One World Services

[smile]

Hi,
Does anyone know if there is a certain point in the UNIX initialization that JDE services should be brought down and brought up. Since every end and start initialization has a number which tells it at what point to begin, we were wondering if there is one that is associated with the JDE RUN and END scripts. We are wanting to automate the process so that services are properly brought down and up if the server is re-booted.

Cleola,

You'll need to evaluate your environment to determine the correct placement of your JDE services in startup and shutdown processes.

Our HP-UX box is used only for our Oracle database and OneWorld Enterprise Server.
In our rc1.d, we have a K087oneworld link, which is the first one to run at that level, followed next by a K090oracle link, which will shutdown Oracle.
In our rc2.d, we have a S990oracle link, which will startup our Oracle database (though we had to change the startup script due to a bug in Oracle's provided startup script at our version). We used to have a S995oneworld link, but we never could get OneWorld to startup at time of reboot, so we renamed this link to s995oneworld (note the lowercase s). So after every reboot, we have to manually start OneWorld.
The S990 and S995 are the last two things to run in the rc2.d directory.

Fortunately, our HP-UX box is very stable, and we have a good UPS backup system, so we don't get many unexpected reboots.

Hope this helps.

My box is a Solaris and all we run on it is the JD Edwards we have a separate Database server. I don't know much about the UNIX shutdown process and the UNIX admin does not know much about the JDE process so you see my dilemma

We generally don't have a lot of re-boots but for the occasional one I would like services to go down properly and come up properly.

Cleola -

The JDE services should be started and stopped at the run level you use for
your other applications. Typically, that will be a multiuser run level like
2 or 3. Look at /etc/rc or sbin/init.d or the man pages for more info.

We use run level 3.

Regards,
Tony

Cleola -

The scripts you would use to start and stop OneWorld based on the UNIX
start/shutdown process are similar to any other unix start/stop scripts
(also called rc scripts) for any other application on the box. If you do a
search on the web, you would be able to find some examples. Then, you
basically tailor them to include the RunOneWorld.sh and EndOneWorld.sh
scripts plus whatever else you want (notifications, etc). Make sure to use
a multiuser run level (like 2 or 3).

The UNIX admin should know how to create rc scripts. If he does not, a web
search will do the trick, man pages ("man init" is a starting point), the
system manuals, or tell the admin to go purchase a UNIX Admin book. This is
a basic UNIX sys admin function. JDE did all the work in creating the
RunOneWorld.sh and EndOneWorld.sh scripts. The rest is just wrapping it in
a format suitable to Solaris.

Tony

Re: RE: UNIX Auto Shut Down and Start Up of One World Services

We are doing it that way using 3 and it is being started by the root account but when it tries to start services it only brings up one process and then it gives an IPC onfilict. It creates a jde.log and a debug.log file in etc/rc2.d, here is what the jde.log says

6613 Thu Aug 14 08:33:04 2003 netsock.c394
2-getservbyname returned No such file or directory

6613 Thu Aug 14 08:35:29 2003 netsig.c513
Net program ended, pid = 6613, signal = 2

6613 Thu Aug 14 08:35:29 2003 ipcpub.c3622
API ipcSawZombieProcV1 : process 6613 set to Zombie in entry 0

for now I am starting it manually but would like it to be in the script in cans I am not around and the server goes down or gets re-booted or something.