I have still not fixed the issues that we are having on our servers, but I have a temporary fix in place which gives the users un-interrupted usage during the business day.
At the moment I am having to re-start the servers at 0630 and 1930 EVERY day that's including Saturday and Sunday. Up until now I had not set a scheduled task to re-boot the servers as one of the servers needs to then be logged on and applications run.
There is also another PC that is logged on and runs something that is called a grs queue. The way the applications work is server1 is the SQL server, Server2 is the application server for the Database it provides a userapps$ share that hold the exe's and the dll's that are required for all the web apps to run. It also provides the Web Menu for the system. Server2 needs to be logged on as an event monitoring user which runs outlook, the event server and event scheduler. And then there is PC1. PC1 is then logged on as a queue user which monitors the print queue on the application.
Server1 and Server2 are the two servers that are having the issues where every 14 hrs it looses its ability to talk to other servers on the network. (this issue is really bugging me now as every fix I have come across has not worked at all).
The problem is when I reboot Server1 the applications that are running on Server2 and PC1 start complaining as there is no SQL connection. Once I have re-booted Server2, logged on and start the applications the server work as normal. However PC1 is still throwing up errors about loosing connectivity to the SQL DB but everything else on the server is running ok.
So here is my answer:
Firstly I set Server2 to use the auto logon feature on Windows. see the following MS article to do that:
http://support.microsoft.com/kb/234562
once I had enabled the auto login I then created a script on one of my Windows 2003 Server's to remotely restart Server1 then wait 2 minutes and then restart Server2 every morning at 0630 and 1930 the way I did that was like this:
shutdown /r /m \\Server1 /t 10
ping localhost -n 120 > nul
shutdown /r /m \\Server2 /t 10
The reason I have put the ping in there is that it will send itself 120 ping requests (which work out to be 120 seconds or 2 min).
The reason I wanted it to wait was the SQL server (Server1) needs to be up before Server2 is up.
Once the 2 servers re up Server2 logs on automatically and then in the start I I than put a batch file that starts outlook, waits a few seconds and then starts the event server, waits a few more seconds and then starts the Event Scheduler.
For PC1 I created a batch that used PSKILL.EXE from PC Tools. This kills the proces grsproc.exe and then waits 8 minutes (again enough time for the SQL server to come back up on line and then starts the process again. This again was set to run at 0630 and 1930 every day.
So this should then save me having to wake up early, restart the servers and and make sure I am home for the evening re-start. Hopefully tomorrow will bring in some positive news!!!
Obviously this has not fixed the issue which is still really annoy as there is no real reason for the servers having network connectivity issues every 14 hrs.
