Full volume dumps occur once/week and incremental backups are taken nightly.
Week1 Week2
i.e. DIIIIII DIIIIII (D=Full Volume DUMP, I=INCREMENTAL)
A recovery test was performed with week 2 day 3 as the recovery date, yet
during DFHSM volume recovery,
DFHSM generated individual dataset recover commands for data created *AND*
migrated prior to the dump for Week2.
i.e. the data was *not* on the volume as of the expected dates involved.
(Week2 Day1-Day3).
Reading in the DFSMShsm Storage Administration Guide( SC23-6871-02), under the
topic
"Full-volume restore with update" (pp 317 - 320) I find:
"DFSMShsm creates a list of candidate data sets from either or both the dump
and backup VTOC
copy data sets to control the volume recovery process." (pp 318).
It is not clear which vtoc/dumpvtoc copies are included.
I would have expected the dumpvtoc copies from the 2nd week and the appropriate
vtoc copies for the incrementals to be used in the
recovery process, but it seems that the dumpvtoc copies from week 1 were also
included when preparing the candidate list.
I have verified that no vtoc copies from week 1 existed at the time of the
recovery.
Can anyone confirm or refute that *both* dumpvtoc copies would be used in this
scenario?
The DUMPCLASS definition used is as follows:
DEFINE DUMPCLASS( /* DEFINE A DUMPCLASS */+
xxxxxx /* NAME OF DUMPCLASS */+
DISPOSITION('USER CRITICAL') /*IDENTIFY DUMPS FOR OFFSITE */+
DAY(0)AUTOREUSE NODATASETRESTORE FREQUENCY(7) +
NORESET RETPD(15) STACK(100) UNIT(FAST) VTOCCOPIES(2))
The RECOVER commands issued were as follows:
HSEND RECOVER * TOVOLUME(xxxxxx) UNIT(3390) -
REPLACE FROMDUMP(APPLYINCREMENTAL)
Thanks in advance for you time and attention,
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
