RE: Server Hung and BRMS Control Group

["Ingvaldson, Scott" <Scott.Ingvaldson@xxxxxxxxxx>]

I like to run a "shutdown" program, similar to the Startup program, before we issue the ENDSBS SBS(*ALL) OPTION(*IMMED).  This ensures that anything that needs to be shut down gracefully or needs more than a minute to end gets a head start.

Currently in our "shutdown" program we're doing:

	CHGSYSVAL  SYSVAL(QSFWERRLOG) VALUE(*NOLOG)
	ENDJOB JOB(QCSTSRCD)
	ENDTCPSVR SERVER(*HTTP)
	ENDTCPSVR SERVER(*MGTC)
	ENDHOSTSVR SERVER(*ALL)
	ENDTCP

and some other things specific to our system.

Good candidates are Domino servers, WebSphere servers, MQ Series and of course any proprietary application you are running.  

This keeps the nasties to a minimum and makes the ENDSBS *ALL run quickly.  In the long run this seems to solve a lot of the unspecified "Integrity Problems" and makes IPLs go a lot more smoothly.


Regards,
 
Scott Ingvaldson
Senior IBM Support Specialist
Midwest Region Data Center
Fiserv.


-----Original Message-----
From: rob@xxxxxxxxx [mailto:rob@xxxxxxxxx] 
Sent: Monday, April 19, 2010 11:04 AM
To: Midrange Systems Technical Discussion
Subject: Re: Server Hung and BRMS Control Group

Jim,

IBM really may suggest ensuring that TCP/IP is down prior to the ENDSBS. 
I'm all for exhausting PMR's until an acceptable situation is obtained. 
However, it's a document fact now that on supported releases of the OS you 
really should drop TCP/IP prior to the ENDSBS or nasties may occur.


Rob Berendt