At 17:15 -0500 on 03/31/2007, Joel C. Ewing wrote about Re: 3490 off
maintenance June 30 2007:
However, you do need to understand that stacking files needed for
recovery purposes on a tape already containing recovery data does
put the previous data at greater risk. Mechanical drives and
physical tapes have the potential at each tape mount for a failure
that can result in loss of all data on the tape, and if you do
enough stacking on a regular basis it is inevitable you will
eventually have such a failure. Our long-term experience with DFHSM
and CA-Vtape on 3590E drives confirms this exposure.
If you do stack datasets in this fashion you want to be sure there
is a duplexed copy of the data somewhere (possibly on another
stacked tape) that you can fall back to, just to cover that
eventuality and prevent data loss.
All valid points. Although I did not state it in my comments (for
simplicity purposes), if I were implementing such a system, I'd start
out with a new Stack Volume daily (The original comment had to do
with doing a swap every 30 minutes) so you'd get only 48 files per
daily volume and start a new volume each day as well as going to
parallel output in the DCB (or using two DCBs) so you get two (or
more) copies at the same time. Both methods close the window of
failure (by having multiple copies as well as minimizing the number
of times you are appending files to any specific volume).
----------------------------------------------------------------------
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
