Re: Server Hung and BRMS Control Group

[Jim Oberholtzer <midrangel@xxxxxxxxxx>]

One other thing to consider.  Make sure you go into the BRMS System 
and/or Back up policies and get those properly set up.  Often times I 
find folks that are not backing up significant portions of the system 
due to some of the defaults in the policies.  Also the IPL settings in 
the policies and if you modify them the control group, can have a 
significant effect on the way the system shuts down for a restricted 
state save.

Rather than have the *EXIT in the back up control group, to shut down 
IP, it would be better go get to the root cause of why the TCP jobs are 
not ending properly.  It is my contention that while this approach may 
very well work, it is a band aid that will eventually come back to cause 
troubles.  I would reopen the PMR and get the TCP guys on it to find out 
what's really going on.

Jim Oberholtzer
CEO/Chief Technical Architect
Agile Technology Architects, LLC


On 4/18/2010 2:21 AM, Sneha Verma wrote:
> *This message was transferred with a trial version of CommuniGate(r) Pro*
> Thanks Rob.
>   
> The was a TCP/IP job called QSYSWRK in sbs QSYSWRK which did not end ,causing the problem.IBM recommends putting an EXIT program just prior to start of backup that will do the follwoing:
>   
> ENDTCPSVR  
> MONMSG MSGID(CPF0000)
> DLYJOB DLY(30)
>   
> ENDHOSTSVR  
> MONMSG MSGID(CPF0000)
> DLYJOB DLY(30)
>   
> ENDTCP      
> MONMSG MSGID(CPF0000)
> DLYJOB DLY(30)
>   
> We have the related BRMS backup control group configured as follows:
>        Backup                                                        
> Seq   Items       Exit command                                      
>                                                                      
>    10  *EXIT       SNDMSG MSG('Backup will begin now .. Please sign o
>    20  *SAVSYS                                                       
>    30  *IBM                                                          
>    40  *ALLUSR                                                       
>    50  *ALLDLO                                                       
>    60  *LINK                                                         
>    70  *EXIT       CALL PGM(CDSLIB/CC1981UC)                         
>    80  *EXIT       SBMJOB CMD(CALL PGM(QGPL/MSG2)) JOB(MSG2) JO
>   
> Can I insert the the IBM recommended commands between sequence number 10 and 20.
>   
> Off course,I will compile a CL (containing the commands) and insert that CL between sequence number 10 and 20.After insertion,I will change the sequence numbers for the commands starting from *SAVSYS and below.
>   
> Also is the delayjob parameter of 30 seconds or each of the commands ENDTCPSVR/ENDHOSTSVR and ENDTCP ok.
>   
> Please advice asap as I have to schedule this to avoid further issues
>   
> Regards
> Sneha
>
>
> --- On Fri, 16/4/10, rob@xxxxxxxxx<rob@xxxxxxxxx>  wrote:
>
>
> From: rob@xxxxxxxxx<rob@xxxxxxxxx>
> Subject: Re: Server Hung
> To: "Midrange Systems Technical Discussion"<midrange-l@xxxxxxxxxxxx>
> Date: Friday, 16 April, 2010, 2:53 PM
>
>
> What job was hung in QSYSWRK?
>
> This is what I fear.  The job would not end with ENDJOB, ENDSBS or any
> such combination.  Your remaining alternatives are ENDJOBABN or IPL.  In
> general, ENDJOBABN should not be something you code into a CL program to
> automatically try 5 minutes after the failure of ENDJOB or ENDSBS.  Note:
> Read the help on ENDJOBABN.  It does require a 5 minute delay between
> ENDJOB and ENDJOBABN and will enforce that.
>
> If you find that job, then a google search on it may show you something.
> Could be anything from:
> get this ptf
> to
> do this function to end the job politely prior to shutdown
>
> Time to start thinking about moving to a supported version of the OS.
>
>
> Rob Berendt
>