R.S.,

For *SIMPLE* solution that is also *clean* I recommend issuing -STOP DB2 
MODE(FORCE) command for each DB2 subsystem. (Command syntax changes based 
on if you are data sharing or not.)  For tasks not currently in DB2 it 
will post an exit; for tasks in DB2 it will stop during next suspend or 
commit. See DB2 Command Reference for more information.

If you wish more sophistication, you can issue:

        -STOP DDF MODE(QUIESCE)
        *wait n seconds*
        -STOP DDF MODE(FORCE)
        -STOP DB2 MODE(QUIESCE)
        *wait m seconds*
        -STOP DB2 MODE(FORCE)
        *wait p seconds*
        MODIFY <irlmproc-name>,ABEND,NODUMP

for slower (but more orderly) shutdown. "n", "m", "p" vary depending upon 
your experience with DDF threads, long-running jobs, commit frequencies, 
and so forth. Last MODIFY command will quickly bring DB2 down, but it is 
usually reserved as a last resort.

And, for even more sophistication, insert -DISPLAY commands for UTILITY, 
DDF DETAIL, various IRLM-related console commands (such as MODIFY 
<irlmproc-name>,STATUS,ALLD ), and so forth.  These give you more 
information about ongoing processes, hence more info on difficulty of 
restart.

However, even after a hard *crash*, DB2 is built to self-recover all 
ongoing work based on log records. Upon restart at your DR facility, DB2 
should analyze the logs, finalize all committed transaction data to disk, 
backout all in-flight transactions, and report any in-doubt threads on the 
console (and xxxxMSTR JES Log).

Lock Lyon
Compuware Corp

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 

[ [<--- snip --->] ]

The procedure have to be as quick as possible and *SIMPLE*.

Back to the question, what about DB2, which I'm most worried ?
I'd like to "quiesce" all the system activity as much as possible, just 
to avoid long time for get the DR system ready. I don't want to spent 
hours to backout uncommited transactions or fight with JES2 to start it 
up. I also don't want to loose the data, from DB2 (obvoius) and possibly 
from JES2 spool.

-- 
Radoslaw Skorupka
Lodz, Poland

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Reply via email to