Regular Reboot of Lawson System

Where on UNIX, we never do periodic reboots. It seems to be an inherit problem with Windows, reboots and Hangs. Any application on Windows seems to need to be rebooted every so often. Our ADAM/MKS Window server, Citrix farm, print servers, etc need periodic reboots, especially when they hang.

We are also a Windows 2008 shop and I have scripts that shut down all WebSphere and Lawson services, do a clean-up of temp files and logs, and then reboot the servers. My scripts run evey Sunday morning when user activity is next to none. All of my Lawson and WebSphere products are set up as services that automatically restart in a specific order after the server has come back up. Seems to work pretty well for us.

We're on Windows 2003 (R2) and we also do a regular shutdoen of all WebSphere services and full reboot each sunday after we do the full backup of our Lawson production server. We used to have a fairly frequent amount of stability issues and a reboot was the only fix. Since getting up to 9.0.0.5 in March (and getting Websphere up to 6.1.0.27), we've been quite stable though. However, we have continued to do the full reboot on a weekly basis just to be safe.

It depends on your transaction volume. I think your GSC meant by reboot is to clean up work files and log files. Periodically, you should delete the GLTDISKxxxx.idx/dat and TMPxxxxxxx.idx/dat files under your ///work folder and purge the log files in your / also. I would check your work directory if you have not and see how many files you already have (i clean up this directory whenever it's over 5000 files and backup the log files and purge it)

However, do be VERY CAREFUL about deleting work files.
If you have any queued jobs running or in 'needs recovery', you probably shouldn't be rebooting ....

I am confused about rebooting. If a job is in needs recovery, it is not running and is waiting for repairs or changes to be made and then to be re-submitted. How does this affect the determination to do a reboot?
We are on Windows 2003 R2 SP2, WAS 6.1.0.27, 9.0.1.4 Portal & Tech and have had to re-start services for WAS and Lawson during past couple of weeks, in prime time, and are now setup for a once a week reboot, early Monday morning.
What items should we be on the lookout for, prior to the reboot, and the proper repair techniques to follow?

What I meant was: you should never script a blind delete of the work (or any, IMHO) files in conjunction with a reboot. Often what you might want to analyze after the reboot would then be missing (one of the reasons I never suggest removing log files in conjunction with a shutdown/reboot, etc. either). As for the perils of rebooting while a job is in needs recovery, it would depend on 1) why the job is in recovery, 2) how serious the job in recovery is (i.e. is it PR140?) and 3) why you are rebooting.

I should add this. When purging the GLT/TMP files in work directory, you should be really careful and do the following steps:

1) dbdump ckpoint
2) look at the restart-info column in your CKPOINT for any GLTxxxxxx.xxx or TMPxxxxxx.xxx, ***YOU DO NOT DELETE THESE FILES***
3) delete any other GLTXXXXXX.XXX and TMPXXXXX.XXX files in the \work that are more than one day old and not the one showing in your CKPOINT restart-info column.